• Beebe, R. G., , and Bates F. C. , 1955: A mechanism for assisting in the release of convective instability. Mon. Wea. Rev., 83 , 110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bradley, A. A., , and Smith J. A. , 1994: The hydrometeorological environment of extreme rainstorms in the southern plains of the United States. J. Appl. Meteor., 33 , 14181431.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Branick, M. L., , Vitale F. , , Lai C. C. , , and Bosart L. F. , 1988: The synoptic and subsynoptic structure of a long-lived severe convective system. Mon. Wea. Rev., 116 , 13351370.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , and Stensrud D. J. , 2000: Climatology of heavy rain events in the United States from hourly precipitation observations. Mon. Wea. Rev., 128 , 11941201.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Byers, H. R., , and Braham R. R. Jr., 1949: The Thunderstorm: Report of the Thunderstorm Project. U.S. Government Printing Office, 300 pp.

    • Search Google Scholar
    • Export Citation
  • Chumchean, S., , Seed A. , , and Sharma A. , 2004: Application of scaling in radar reflectivity for correcting range-dependent bias in climatological radar rainfall estimates. J. Atmos. Oceanic Technol., 21 , 15451555.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Collier, C. G., 1996: Applications of Weather Radar System: A Guide to Uses of Radar in Meteorology and Hydrology. John Wiley and Sons, 294 pp.

    • Search Google Scholar
    • Export Citation
  • Colman, B. R., 1984: Deep convection occurring above a stable planetary boundary layer. Ph.D. thesis, Massachusetts Institute of Technology, 243 pp.

  • Colorado State University, cited. 2009: NWS DIFAX weather map archive. [Available online at http://archive.atmos.colostate.edu.

  • Crook, N. A., , Carbone R. E. , , Moncrieff M. W. , , and Conway J. W. , 1990: The generation and propagation of a nocturnal squall line. Part II: Numerical simulations. Mon. Wea. Rev., 118 , 5066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. S., 2001: Flash flood forecast and detection methods. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 481–525.

    • Search Google Scholar
    • Export Citation
  • Doswell C. A. III, , , Brooks H. E. , , and Maddox R. A. , 1996: Flash flood forecasting: An ingredients-based methodology. Wea. Forecasting, 11 , 560581.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fabry, F., , Bellon A. , , Duncan M. R. , , and Austin G. L. , 1994: High resolution rainfall measurements by radar for very small basins: The sampling problem re-examined. J. Hydrol., 161 , 428.

    • Search Google Scholar
    • Export Citation
  • Fulton, R. A., , Breidenback J. P. , , Seo D-J. , , and Miller D. A. , 1998: The WSR-88D rainfall algorithm. Wea. Forecasting, 13 , 377395.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Heideman, K. F., , and Fritsch J. M. , 1988: Forcing mechanisms and other characteristics of significant summertime precipitation. Wea. Forecasting, 3 , 115130.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hildebrand, P. H., 1978: Iterative correction for attenuation of 5 cm radar in rain. J. Appl. Meteor., 17 , 508514.

  • Junker, N. W., , Schneider R. S. , , and Fauver S. L. , 1999: A study of heavy rainfall events during the Great Midwest Flood of 1993. Wea. Forecasting, 14 , 701712.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Key, J. R., , and Chan A. C. K. , 1999: Multidecadal global and regional trends in 1000 mb and 500 mb cyclone frequencies. Geophys. Res. Lett., 26 , 20532056.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Konrad C. E. II, , 2001: The most extreme precipitation events over the eastern United States from 1950 to 1996: Considerations of scale. J. Hydrometeor., 2 , 309325.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maddox, R. A., , and Doswell C. A. III, 1982: An examination of jet stream configurations, 500 mb vorticity advection and low-level thermal advection patterns during extended periods of intense convection. Mon. Wea. Rev., 110 , 184197.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maddox, R. A., , Chappel C. F. , , and Hoxit L. R. , 1979: Synoptic and meso-α scale aspects of flash flood events. Bull. Amer. Meteor. Soc., 60 , 115123.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, R. C., 1967: Notes on analysis and severe storm forecasting procedures of the Military Weather Warning Center. AWS Tech. Rep. 200, 94 pp. [Available from Headquarters Air Weather Service, Scott AFB, IL 62225].

    • Search Google Scholar
    • Export Citation
  • Moore, J. T., , Glass F. H. , , Graves C. E. , , Rochette S. M. , , and Singer M. J. , 2003: The environment of warm-season elevated thunderstorms associated with heavy rainfall over the central United States. Wea. Forecasting, 18 , 861878.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1950: The structure and mechanism of the prefrontal squall line. J. Meteor., 7 , 210222.

  • Newton, C. W., 1963: Dynamics of severe convective storms. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 33–58.

  • Niziol, T. A., , Snyder W. R. , , and Waldstreicher J. S. , 1995: Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow. Wea. Forecasting, 10 , 6177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NWS, cited. 2007: WSR-88D rainfall estimation. HL publications and papers. [Available online at http://www.nws.noaa.gov/oh/hrl/papers/papers.htm#wsr88d.

    • Search Google Scholar
    • Export Citation
  • NWS, cited. 2009: National Weather Service glossary. [Available online at http://forecast.weather.gov/glossary.php?word=precipitable+water.

    • Search Google Scholar
    • Export Citation
  • Rhea, O. J., 1966: A study of thunderstorm formation along dry lines. J. Appl. Meteor., 5 , 5863.

  • Riehl, H., and Coauthors, 1952: Forecasting in the Middle Latitudes. Meteor. Monogr., No. 5, Amer. Meteor. Soc., 80 pp.

  • Roebber, P. J., , and Eise J. , 2001: The 21 June 1997 flood: Storm-scale simulations and implications for operational forecasting. Wea. Forecasting, 16 , 197218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rose, S. F., , Hobbs P. V. , , Locatelli J. D. , , and Stoelinga M. T. , 2004: A 10-yr climatology relating the locations of reported tornadoes to the quadrants of upper-level jet streaks. Wea. Forecasting, 19 , 301309.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , Atlas D. , , and Short D. A. , 1990: The estimation of rainfall by area integrals. Part 2. The height–area rain threshold (HART) method. J. Geophys. Res., 95 , 21612176.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ross, B. B., 1987: The role of low-level convergence and latent heating in a simulation of observed squall line formation. Mon. Wea. Rev., 115 , 22982321.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rotunno, R., , Klemp J. B. , , and Weisman M. L. , 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463485.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schultz, P., 1989: Relationships of several stability indices to convective weather events in northeast Colorado. Wea. Forecasting, 4 , 7380.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , and Johnson R. H. , 2005: Organization and environmental properties of extreme-rain-producing mesoscale convective systems. Mon. Wea. Rev., 133 , 961976.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , and Johnson R. H. , 2006: Characteristics of U.S. extreme rain events during 1999–2003. Wea. Forecasting, 21 , 6985.

  • Seo, D-J., 1998: Real-time estimation of rainfall fields using radar rainfall and rain gauge data. J. Hydrol., 208 , 2536.

  • Short, D. A., , Kozu T. , , and Nakamura K. , 1990: Rainrate and raindrop size distribution observations in Darwin, Australia. Proc. URSI-F Open Symp. on Regional Factors in Predicting Radiowave Attenuation due to Rain, Rio de Janeiro, Brazil, Rutherford Appleton Laboratory and Empresa Brasileira de Telecomunicaoes, 35–40.

    • Search Google Scholar
    • Export Citation
  • Smith, J. A., , and Krajewski W. F. , 1991: Estimation of the mean field bias of radar rainfall estimates. J. Appl. Meteor., 30 , 397412.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, E. N., , Hollinger S. E. , , and Kunkel K. E. , 2005: Use of real-time multisensor data to assess the relationship of normalized corn yield with monthly rainfall and heat stress across the central United States. J. Appl. Meteor., 44 , 16671676.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • White, G. H., 1982: An observational study of the Northern Hemisphere extratropical summertime global circulation. J. Atmos. Sci., 39 , 2440.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    The study region is shaded.

  • View in gallery

    Multisensor precipitation estimates from radar rainfall data for the data range 1–31 mm h−1. (top) Histogram of all hourly rainfall rates for all good data in the study area. Any data points with hourly rainfall rates greater than 31 mm h−1 have been removed. (bottom) Cumulative total of all hourly rainfall rates, showing that the 99th percentile occurs at 16 mm h−1.

  • View in gallery

    Example CULL SCEPT event from 2 Apr 2005, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) the surface analysis provided by the Colorado State University DIFAX archive.

  • View in gallery

    CULL event locations.

  • View in gallery

    Median times during which training occurred for all CULL, ULT, and 850TL events.

  • View in gallery

    CULL events indicating the (a) location of the 500-hPa low, (b) location of the 300-hPa low, (c) location of the 850-hPa low, and (d) location of the surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

  • View in gallery

    CULL events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

  • View in gallery

    Example ULT SCEPT event from 2 May 2004, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) the surface analysis provided by the Colorado State University DIFAX archive.

  • View in gallery

    ULT event locations.

  • View in gallery

    ULT events indicating the (a) location of the 850-hPa low and (b) surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

  • View in gallery

    ULT events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) the location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

  • View in gallery

    Example 850TL SCEPT event from 12 Jul 2004, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) at the surface analysis provided by the Colorado State University DIFAX archive.

  • View in gallery

    850TL event locations.

  • View in gallery

    850TL events indicating the (a) location of the 850-hPa low and (b) surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

  • View in gallery

    850TL events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 22 22 3
PDF Downloads 11 11 0

Synoptic Environments Associated with the Training of Convective Cells

View More View Less
  • 1 Department of Geosciences, Mississippi State University, Mississippi State, Mississippi
© Get Permissions
Full access

Abstract

Many studies have been done on synoptically forced systems and heavy rainfall, but little research has gone into the forecasting of training convective storms specifically in a synoptically forced environment. Training convective storms are recognized as the propagation of convective cells repeatedly over the same location. The research in this paper examines 38 separate synoptically forced convective extreme precipitation training (SCEPT) events to find trends and consistencies in their synoptic environments. Three separate cases were found in which a SCEPT event occurred: closed upper-level trough (CULL), upper-level trough (ULT), and 850-hPa trough–low (850TL). Each event occurred in areas of precipitable water greater than 36.42 mm (1.43 in.), near maximums of 850-hPa moisture convergence and 700-hPa upward vertical velocities, under the 850-hPa jet, and in the warm sector of a midlatitude cyclone. CULL and ULT events occurred in strongly forced synoptic environments where 500- and 300-hPa troughs were evident and generally positively tilted. Little upper-level forcing, above 700 hPa, was found in 850TL events.

Corresponding author address: Jamie L. Dyer, P.O. Box 5448, Mississippi State, MS 39762-5448. Email: jamie.dyer@msstate.edu

Abstract

Many studies have been done on synoptically forced systems and heavy rainfall, but little research has gone into the forecasting of training convective storms specifically in a synoptically forced environment. Training convective storms are recognized as the propagation of convective cells repeatedly over the same location. The research in this paper examines 38 separate synoptically forced convective extreme precipitation training (SCEPT) events to find trends and consistencies in their synoptic environments. Three separate cases were found in which a SCEPT event occurred: closed upper-level trough (CULL), upper-level trough (ULT), and 850-hPa trough–low (850TL). Each event occurred in areas of precipitable water greater than 36.42 mm (1.43 in.), near maximums of 850-hPa moisture convergence and 700-hPa upward vertical velocities, under the 850-hPa jet, and in the warm sector of a midlatitude cyclone. CULL and ULT events occurred in strongly forced synoptic environments where 500- and 300-hPa troughs were evident and generally positively tilted. Little upper-level forcing, above 700 hPa, was found in 850TL events.

Corresponding author address: Jamie L. Dyer, P.O. Box 5448, Mississippi State, MS 39762-5448. Email: jamie.dyer@msstate.edu

1. Introduction

Extreme rainfall events often lead to flash flooding and pose a threat to both life and property (Brooks and Stensrud 2000). Increased urbanization has amplified the risk of flash flooding by decreasing infiltration rates through the increase of nonporous ground cover (i.e., asphalt, concrete). This results in greater runoff over a shorter time period, leading to an increased flood risk. Agriculture is also threatened by extreme rainfall, as the rains can wash away soils and destroy plants during a flooding event. Thus, it is important that extreme rainfall events be better investigated to help in forecasting, resulting in saved lives and reduced property damage.

Extreme rain events can be separated into five general types: mesoscale convective systems (MCSs), tropical, terrain forced, high-precipitation supercells, and synoptic (Schumacher and Johnson 2005). Of these five categories, synoptically forced extreme rainfall has been most frequently investigated (Newton 1950; Branick et al. 1988; Rotunno et al. 1988). However, the dynamics resulting in synoptically forced training convective rainfall have not been studied in as much detail (Moore et al. 2003; Junker et al. 1999). This research gap indicates that further study is needed to investigate synoptically forced convective extreme precipitation training (SCEPT) events in order to better forecast such occurrences in the future.

High rainfall totals often occur when convective cells organize into an “echo training” event recognized as the movement of convective echo returns on radar over the same location (Doswell et al. 1996; Davis 2001). As the storm motion becomes tangent to the line of storms, an increase in the total rainfall occurs. A training convective system is a “process of subprocesses (convective cells)” where the negative propagation or stationary movement of the convective cells can negate the forward motion of the entire convective system (Doswell et al. 1996). The propagation of convective cells over the same location describes a training event; however, a precise definition of a training event, in regard to duration and precipitation amount, is not stated in the previous literature. Rather, past research has used 50-yr recurrence intervals (Schumacher and Johnson 2005, 2006) and/or 24- to 48-h total rainfall amounts (Konrad 2001; Moore et al. 2003) to define an extreme precipitation event. This methodology does not specify a length of time of the event itself or how much precipitation fell during the period of heaviest rainfall.

This study will first define a SCEPT event in regard to duration, convective precipitation rate, and total precipitation. Doswell et al. (1996) states that the total storm precipitation is proportional to both the rainfall rate and duration at any point; therefore, it is important to examine how storm characteristics (i.e., moisture availability, precipitation efficiency, etc.) differ from one storm system to another. To accomplish this, events will be found that satisfy the established definition based on multisensor precipitation estimates. Select atmospheric variables will then be examined in and around each SCEPT event location using the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) North American Regional Reanalysis (NARR) dataset. Variables include precipitable water, lower- and upper-level winds, temperature advection, upper-level vertical velocities, and the locations of jet streaks, as these can all affect storm strength, movement, and precipitation rate in a synoptically forced environment.

The goal of this research is to define the synoptic conditions necessary for the generation of a SCEPT event and to establish a definition of an extreme precipitation training event in terms of its duration and precipitation rate. With this information, forecasters will be able to recognize that a life-threatening SCEPT event could potentially occur in their forecast region. As stated earlier, SCEPT events have been greatly underresearched in the literature despite the abundance of research done on synoptically forced systems in general and the training of convective cells by mesoscale convective systems. This paper will expand on past training research to broaden our understanding of the development of SCEPT events.

2. Synoptically forced systems

Synoptically forced systems have been proven to produce extreme rainfall (Maddox et al. 1979; Heideman and Fritsch 1988; Bradley and Smith 1994; Branick et al. 1988). Organized convective systems, with embedded severe weather and associated heavy rainfall, are often linked with frontal boundaries that help focus convection (Maddox et al. 1979: Newton 1963; Rhea 1966; Branick et al. 1988). Deep convection develops in a strongly sheared baroclinic environment where differential cyclonic vorticity advection in the middle and upper troposphere, in advance of a short-wave trough or jet streak, is involved in organizing large-scale ascent (Riehl et al. 1952; Beebe and Bates 1955; Miller 1967; Branick et al. 1988).

Maddox et al. (1979) defined two heavy rainfall types specific to this research, synoptic and frontal, during research of heavy rainfall events in the eastern United States. A synoptic event develops primarily in the transitional seasons and is composed of a strong 500-hPa trough moving slowly eastward. Heavy rainfall during a synoptic-type event occurs in the warm sector out ahead of a slow-moving or stalled frontal boundary. A frontal event occurs when a warm front stalls in an east–west fashion and warm-air advection, generally from the south, overruns the front resulting in heavy rainfall on the cool side of the boundary. These events are most common during the night in the summer months.

Heideman and Fritsch (1988) found 50% of extreme rainfall events, though not necessarily training events, were synoptically forced (non–mesoscale convective systems) while 80% of the synoptically forced events were produced through convection in the eastern United States during the warm season. Synoptically forced events in their study were defined as events that occurred near a frontal boundary. Schumacher and Johnson (2006) found 93% of synoptic events were convective, but only 25% of the heavy-rain events in their study were primarily of synoptic forcing and normally occurred during the cool season in the eastern United States.

A pronounced seasonal difference in atmospheric conditions exists between the warm (summer) and cool (winter) seasons across the globe (White 1982). White (1982) stated that standing Rosby waves are smaller in scale during the summer than in the winter, or less amplified. This has a direct impact on this study considering that midlatitude cyclones are commonly associated with upper-level troughs. Additionally, the total solar radiation increases in the Northern Hemisphere during the summer months, resulting in increased heat (White 1982) at the surface and aloft, which leads to the strongest westerly tropospheric winds shifting north toward greater thermal gradients.

A statistically significant increase in the number of 500-hPa cyclones (synoptically driven) occurred from 1958 to 1997 (Key and Chan 1999). This again indicates the increasing importance of studying SCEPT events, as these types of events are becoming more common. Konrad (2001) found the vast majority of 500-hPa lows were primarily 500–1500 km northwest of a heavy rain event in the eastern United States. The study also found the low to track east-southeasterly while deepening. Maddox et al. (1979) found the 500-hPa trough to be nearly stationary as the heavy rainfall event occurred.

Moore et al. (2003), during a study of heavy precipitation events from elevated thunderstorms during the warm season in the central United States, found areas of near-neutral to weakly positive differential vorticity advection were most common near areas of heavy precipitation. A previous study by Maddox and Doswell (1982) in which three case studies were examined during the spring months in the central United States agreed with these findings, indicating that regions of 500-hPa differential vorticity advection are not good indicators of extreme precipitation.

The influence of the low-level jet is considered to be a good indicator of when heavy rainfall will occur (Moore et al. 2003; Junker et al. 1999; Roebber and Eise 2001). Buoyancy can be provided by a warm, moist airstream (the low-level jet), which can ascend over a cold boundary layer, resulting in elevated thunderstorms (Colman 1984). The low-level jet has been found to help initiate convection, which can be enhanced by warm-air advection (WAA; Maddox and Doswell 1982), normally oriented south-southwesterly during heavy-rain events (Moore et al. 2003). Junker et al. (1999), in a study of heavy rainfall from June through September in the Upper Midwest during the Great Midwest Flood of 1993, found that a wide low-level jet can lead to a stretched region of moisture convergence, enhancing the potential of merging or training cells, which is backed up by Moore et al. (2003). Both studies also show that the axis of moisture convergence parallels the midlevel flow in training situations. This paper will look at the jet at 850 hPa for consistency in geopotential height since the actual low-level jet can migrate from one vertical level to another depending on the environment.

There are differing results on the location of maximum 850-hPa convergence. Junker et al. (1999) found the maximum convergence occurred approximately 2° (∼220 km) north of the heavy rainfall location, while Moore et al. (2003) found the maximum 850-hPa moisture convergence 200 km south of a mesoscale convective system centroid. This research paper will establish which conclusion, or other possibility, best describes a SCEPT event.

Precipitable water is a measure of the depth of liquid water in a column of the atmosphere if it were completely condensed to the surface (NWS 2008). Precipitable water was at a local maximum at the location of heaviest rainfall during strongly forced synoptic events in which a long-wave trough was present with an intense synoptic-scale frontal system in the south-central plains throughout the year (Bradley and Smith 1994). Further studies have found precipitable water greater than 43.2 mm (1.70 in.) accompanying severe weather and heavy rainfall events in the warm season in northeast Colorado (Schultz 1989) and in the Midwest (Roebber and Eise 2001), while Maddox et al. (1979) found precipitable water values of 37.1 mm (1.46 in.) during synoptically forced heavy rainfall events in the spring and fall east of the Rocky Mountains.

3. Data

a. Multisensor precipitation estimates

Precipitation estimation via radar begins by processing the raw precipitation reflectivity data using a ZR relationship, which relates the reflectivity (Z) to the precipitation rate (R; mm h−1). The data are integrated to produce hourly precipitation estimates and are output as a digital precipitation array (DPA). At this point, the data are referred to as stage I data. Considerable error can occur due to inherent problems with radar precipitation observations, including beam overshoot and attenuation (Hildebrand 1978), ground clutter (Collier 1996), and range-dependent bias (Fabry et al. 1994). Further, ZR relationships are strongly dependent on precipitation type, requiring different ZR relationships for convective, stratiform, and frozen precipitation (Chumchean et al. 2004). Thus, the choice in ZR relationship can be a major influence on later precipitation estimates.

Each radar site is effective to a radius of 230 km, referred to as the radar umbrella. Within this field, a calculated mean field gauge–radar bias, using a Kalman filtering approach, is fit to all grid cell values (Smith and Krajewski 1991), and other corrective local adjustments are made to the radar data. When finished, the radar data are called stage II and provide a multisensor precipitation field that includes both the corrective mean bias and the local gauge-derived biases (NWS 2007).

Before 2003, National Weather Service (NWS) River Forecasting Centers further analyzed the data to create the stage III product, which was a mosaic of all sage II data for the forecasting center’s specific region. In 2003, the stage III algorithm was replaced with the multisensor precipitation estimator (MPE). Within the MPE, a mosaic of stage II data is created, but an additional weighted adjustment is included, based on the surface gauge distance from the precipitation event (Westcott et al. 2005). More weight is given to the radar estimate as the precipitation event occurs farther from a rain gauge, allowing for adjustment based on within-storm variability (Fulton et al. 1998; Seo 1998). Currently, MPE data are archived back to 2003 and come in a nominal 4 km × 4 km resolution.

b. NCEP–NCAR North American Regional Reanalysis

The National Centers for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) began the NCEP–NCAR Global Reanalysis Program in 1991 (Kalnay et al. 1996). The project’s beginnings originated as an outgrowth of the National Meteorological Center’s (NMC, now known as NCEP) Climate Data Assimilation System project and the program was designed to help correct for apparent “climate changes” resulting from adjustments introduced in the NMC operational Global Data Assimilation System in the mid-1980s. The reanalysis program involves the recovery of surface, ship, rawinsonde, aircraft, satellite, and other data sources.

Data are currently available from 1957 to the present and include comprehensive analysis and diagnostic fields presented in synoptic form (every 6 h, at 0000, 0600, 1200, and 1800 UTC) on a 2.5° latitude × 2.5° longitude grid. An offshoot of this program is the North American Region Reanalysis (NARR) project, with data currently available from 1979 to the present for the North American realm. Comprehensive analysis and diagnostic fields are presented in synoptic- and mesoscale forms at a finer time scale (every 3 h) and resolution (32 km). The diagnostic fields included in this study are height fields (300, 500, and 850 hPa), 300-hPa jet streaks, 500-hPa differential vorticity, 700-hPa vertical velocity, 850-hPa temperature, surface moisture convergence, and precipitable water.

4. Methodology

The study area for this research project is shown in Fig. 1. This region was chosen because of its high population density and thus its high flash flooding casualty potential. The states bordering the Gulf of Mexico were excluded because of the increase in tropical influences, which is outside the scope of this study. The Ohio Valley and mid-Atlantic have a good balance of tropical moisture and midlatitude cyclone influences that can facilitate extreme precipitation events. Events in the Ohio Valley are often associated with nocturnal mesoscale convective systems while events along the mid-Atlantic occur primarily in the afternoon and evening during frontal passages (Schumacher and Johnson 2006). MPE algorithm rainfall data are fully archived back to the year 2004; therefore, the time frame for the study was restricted to between 2004 and 2006.

Before beginning the synoptic analysis, it was necessary to obtain a definition of a SCEPT event. To do this, multisensor precipitation estimate radar data were utilized by creating a histogram of each hourly rainfall total for every data point during the 3-yr study period (Fig. 2). The bins were designated at 1 mm h−1 intervals. As can be expected, rainfall rates were most common near 1–2 mm h−1, with instances of higher rainfall decreasing exponentially.

It was found that 99% of all rainfall rates occur between 1 and 16 mm h−1 (Fig. 2). This signifies rainfall rates of 17 mm h−1 and greater were in the 99th percentile of all of the quality controlled multisensor precipitation estimate radar data. By using the 99th percentile, we ensured the event is anomalous, characterized as having an extreme rainfall rate. Past research found that stratiform precipitation events rarely exceeded 10 mm h−1 [0.4 in h−1; Short et al. (1990); Rosenfeld et al. (1990)]. Thus, using 17 mm h−1 also ensured that convective precipitation processes were responsible for the extreme precipitation events used in this study.

To be certain that training was occurring, it was important that high rates of precipitation continued for a sustained period of time. This required that convective cells continued to propagate over a single location. A single-cell thunderstorm’s lifespan averages 75 min (1.25 h; Byers and Braham 1949). To make certain that multiple convective cells propagated over the location, a 3-h time span was used. Thus, if rainfall rates exceeded 17 mm h−1 for at least three consecutive hours, the event was considered an extreme precipitation training event.

Using these criteria, all of the grid points in which rainfall rates were greater than 17 mm h−1 for three consecutive hours were found. All event dates were recorded and then analyzed to determine the type of weather event that caused training to occur. Using archived surface and radar maps (Colorado State University 2009), each event was subjectively determined, similarly to the methodology used by Heideman and Fritsch (1988), to be either synoptic (influenced by a midlatitude cyclone, warm front, cold front, or stationary front), tropical (origination from tropical waters; usually a hurricane, tropical storm, or depression), an MCS not dominated by synoptic processes (no midlatitude cyclone or frontal boundaries present near the MCS), or of airmass (not in any other category) origin (Table 1). If an extreme rainfall event occurred near a frontal system and no tropical influences were noted, it was considered a synoptically forced event and was included in the study. All tropical, MCS, and airmass events were removed from the dataset. Once this manual analysis was completed, 36 SCEPT events were found in the study region from 2004 to 2006.

A central location for each event was then determined. If the event only occurred at one grid point on a particular day, then that grid point was the event location. If multiple grid points on a single day contained a SCEPT event, a central location was determined by examining all gridpoint locations for that day, finding the greatest cluster of data points through visual interpretation, and choosing one grid point to represent that event from the cluster. If two separate areas, or clusters, were found on a single day, the largest cluster was used in this dataset.

To analyze the synoptic conditions of each event date and location, NARR data were utilized. The comprehensive analysis and diagnostic fields presented in synoptic form are available every 3 h during the study period. Each event required at least two time periods to establish trends in the data. It was noted that the closest field to the event time would be required to determine the synoptic conditions at the time of the event. The second field utilized was the previous 3-h diagnostic field.

NARR data for geopotential height (300, 500, and 850 hPa), 500-hPa differential vorticity, 300-hPa jet streaks, upward vertical velocity (UVV) at 700 hPa, 850-hPa temperature fields, and precipitable water were retrieved for each event. For each height field, the trough axis was found and determined to be deepening, weakening, or neutral and either tilted positively, negatively, or neutrally. A deepening (weakening) trough was defined in this study as the geopotential height of the trough decreasing (increasing) [become lower (higher)] and/or the magnitude of the trough increasing (decreasing) [generally driving the trough toward the south (north)]. A positively (negatively) tilted trough was defined as the axis of the trough’s base tilted toward the west (east). A short-wave trough was defined as less than 800 km in amplitude and wavelength, while a long-wave trough was greater than 800 km in amplitude and/or wavelength. Any closed lows at the pressure levels were recorded. Additionally, the axis of greatest precipitable water from the event location to a moisture source (Gulf of Mexico or Atlantic Ocean) was noted, described as the moisture tongue in this paper.

5. Results

Three distinct synoptic situations were found that produced SCEPT events: 1) closed upper-level low (CULL), 2) upper-level trough (ULT), and 3) 850-hPa trough–low (850TL). Each case has its own specific characteristics that normally occur within the synoptic environment; however, a few characteristics were common in all cases that included strong upward vertical velocities found at the 700-hPa level, warm-air advection (WAA), and precipitable water averaging greater than 36 mm (1.4 in.). Each case will be described in detail in this section.

Forty-seven percent of all SCEPT events (17 events; Table 2) were found to have a closed 500-hPa low (CULL) present to the west of the SCEPT event location. An additional nine cases were found where a long-wave 500-hPa trough (ULT) was present but lacked a closed 500-hPa low. This means 72% of all SCEPT events from 2004 to 2006 were of the strong synoptically forced variety with a long-wave trough. The final 28% fell within the 850TL category, indicating a weaker synoptically forced environment without the presence of a long-wave 500-hPa trough.

a. Closed upper-level low (CULL)

Nearly half of the events (47%) found in this study contained a 500-hPa closed low (when using 10 gpm height contour increments) to the west of the event location (Table 2). An example CULL event from 2 April 2005 is depicted in Fig. 3 for data at 300, 500, and 850 hPa, and at the surface. The clear-closed 500-hPa low is centered over southern Illinois–Indiana with a midlatitude cyclone depicted at the surface centered over West Virginia. The SCEPT event on this date developed in far northern Virginia north of the warm front. This specific event could be considered a hybrid of the synoptic and front types described by Maddox et al. (1979). Most of these events (76%) occurred during the fall to spring period from September through April, when the strongest synoptic forcing exists (Table 3). Sixty-five percent of the CULL events occurred in Kentucky, Indiana, and Illinois (Fig. 4).

With regard to time of day, all events except for one occurred between 0700 and 1600 UTC with the majority of events occurring in the morning hours. This was different from what past research had found where the peak of heavy rainfall was from 0300 to 0500 UTC (Schumacher and Johnson 2005; Fig. 5). Crook et al. (1990) found that a delayed second maximum in vertical motion occurred 6 h after the initial vertical motion during afternoon heating. However, this would still only place the secondary maximum near 0600 UTC if convection initiated in the afternoon near 0000 UTC.

In the upper levels of the atmosphere, the 500-hPa closed low was to the northwest of the CULL SCEPT event with a mean distance of approximately 960 km (600 mi, r = 0.04; Table 4) to the west-northwest (Fig. 6a; note the scale differences between variables, which will remain consistent throughout the paper), similar to findings in Konrad (2001). Seventy-six percent of the 500-hPa troughs were positively to neutrally tilted. This deviates from normal observations during severe weather outbreaks, when the trough becomes negatively tilted, thus enhancing vertical uplift by positive differential vorticity advection (Riehl et al. 1952; Beebe and Bates 1955; Miller 1967; Branick et al. 1988). As was stated in section 2, differential vorticity advection has little influence on heavy-rain events (Moore et al. 2003). This research concurs with these findings where 71% of the events had little to no positive differential vorticity advection. The 500-hPa trough was deepening, or near neutral, in 82% of the events. This again indicates that the trough was still strengthening as the CULL SCEPT event occurred.

The 300- and 500-hPa lows were nearly vertically stacked at the time of the events with the 300-hPa low to the west-northwest, approximately 70 km (40 mi) more than the 500-hPa low, or 1030-km event (640 mi, r = 0.01; Fig. 6b, Table 4), from the CULL SCEPT event. The CULL SCEPT events were located on the right side of a jet streak during 76% of the events; however, no preference in entrance or exit regions was found. The presence of a jet streak indicated an increased ability of the upper levels to help evacuate mass, assisting in sustaining the convection. The lack of a consistent quadrant, preferably the right-entrance region, where increased lift has been found to occur (Rose et al. 2004), indicates that a jet streak itself is enough to enhance convection rather than the specific quadrant. The trough tilt at 300 hPa was difficult to determine with only 53% of the events having a positive to near neutrally tilted trough when training occurred. However, the trough was deepening three-fourths of the time, enhancing the jet streak.

The center of the 850-hPa low had a moderately strong linear pattern (r = −0.34, p < 0.10; Table 4) toward the northwest of the CULL SCEPT event location, with an average distance of approximately 665 km (410 mi; Fig. 6c). This position normally placed the 850-hPa jet over the site in which training occurred. WAA was found in most cases (88%). The tilt of the 850-hPa trough was primarily positive or near neutral in 82% of the events. Seventy percent of the CULL SCEPT events occurred when the 850-hPa low–trough deepened or held steady. The combination of a deepening, positively tilted trough indicated a strengthening storm system at the time of the event.

Frontal boundaries were the foci of convergence and convective initiation in past research (Ross 1987; Schumacher and Johnson 2005) and this was found to be the case in this study as well. On average, the cold front was 405 km (250 miles; Table 5) west along a line of equal latitude not influencing the event. However, the warm front was only 95 km (60 mi; Table 5) south along a line of equal longitude from the CULL SCEPT event location on average. This placed the CULL SCEPT event very near the warm front, similar to the frontal type described by Maddox et al. (1979). The midlatitude low location averaged 490 km (305 mi, r = −0.07; Table 4) to the west of the CULL SCEPT event location (Fig. 6d).

Past research provided differing results concerning surface moisture convergence. Junker et al. (1999) found moisture convergence to be greatest 2° north of the heavy rainfall location while Moore et al. (2003) concluded that the greatest convergence was 175 km (120 mi) south of a mesoscale convective system centroid. This study found the greatest surface moisture convergence to be in a linear pattern from the southwest to northeast of the event location, which is the average flow pattern of the 850-hPa jet (Fig. 7a). Instead of being centered north or south of the training event, the maximum surface moisture convergence was along the 850-hPa flow just upstream, over, or downstream of the CULL SCEPT event location in a moderately strong linear pattern (r = 0.45, p < 0.05; Table 4). This indicated that the location of the CULL SCEPT event was not always in the same direction from the greatest surface moisture convergence, but rather that training will occur along the flow near the greatest moisture convergence. The one event that had a maximum surface moisture convergence to the southeast of the CULL SCEPT event, and also one that had a maximum far to the northwest, occurred under a southeasterly 850-hPa flow regime. Half of the maximum locations of moisture convergence had a long axis along the flow and half were perpendicular to the flow, indicating that axis orientation did not specify training was about to occur.

The 850-hPa jet traversed over the CULL SCEPT location during 82% of the events oriented from south-southwest to north-northeast. This orientation brought abundant moisture from the Gulf of Mexico to most of the study region, except along the eastern seaboard due to blocking from the Appalachian Mountains. This likely explains why a maximum of events occurred in the midwestern United States (Fig. 4).

Considerable moisture was in the atmospheric column above the CULL SCEPT event, with precipitable water values averaging 36.2 mm (1.42 in.; Table 5). Recent research found that values of 43.2 mm (1.70 in.) of precipitable water resulted in heavy-rain events, which were often associated with warm season events in the central United States (Schultz 1989; Roebber and Eise 2001). However, the values found for the CULL SCEPT type are very close to the values found in Maddox et al. (1979) for synoptic-forced heavy rainfall in the spring and fall, which resembles the time period in which CULL SCEPT events occurred. These values indicate that heavy rainfall is possible, but additional variables such as the positioning of the upper-level trough, warm front, and surface moisture convergence are needed in order to locate an area in which training extreme rainfall will occur. The moisture tongue of high precipitable water values had a linear pattern (r = 0.31; Table 4) to the northeast, extending 550 km (340 miles) from the CULL SCEPT event (Fig. 7b), while advected from the south-southwest.

At the 700-hPa level, upward vertical velocities (UVVs) averaged −0.54 Pa s−1 at the CULL SCEPT event location. In three-fourths of the events, the highest UVV values were found within 300 km (185 mi) of the site of training at the time of the event. As shown in Fig. 7c, almost all of the greatest UVV values were located either to the west-southwest or the east-northeast of the CULL SCEPT event, producing a linear pattern (r = 0.25). A west-southwest to east-northeast orientation aligned with the flow at 700 hPa.

In summary, a CULL event is defined by a positively tilted, strengthening trough at 850 and 500 hPa. The 300- and 500-hPa lows were nearly vertically stacked, with the 300-hPa low lagging 70 km (40 mi) behind, on average. In the lower levels, the 850-hPa low had a moderately strong linear pattern (r = −0.34, p < 0.10) toward the northwest, averaging approximately 665 km (410 mi) from the event location. The 850-hPa jet traversed over the CULL SCEPT event in most of the cases, helping to advect moisture from the south-southwest. Precipitable water averaged 36.2 mm (1.42 in.), with the maximum extent of the moisture tongue 550 km (340 mi) to the northeast. Surface moisture convergence occurred along the flow of the 850-hPa jet in a moderately linear pattern (r = 0.45, p < 0.05) from southwest to northeast. The 700-hPa UVVs occurred along the flow at 700 hPa as well as from the west-southwest to east-northeast in a linear pattern (r = 0.25). The CULL SCEPT event was located along the warm front of a midlatitude cyclone under warm-air advection during most events. The midlatitude low pressure center was 490 km (305 mi) to the west-northwest (r = −0.07). CULL events occurred between 0700 and 1600 UTC, primarily in the midwestern states during the cool season months.

b. Upper-level trough (ULT)

A quarter of the SCEPT events (nine events) had a 500-hPa trough without a closed low (Table 6). The lack of a closed low indicates the event was not as strongly synoptically forced as the CULL events, which the data in this section support. However, the reduction in the dynamics does not indicate a reduction in rainfall potential. An example of an ULT event from 2 May 2004 is shown in Fig. 8 with geopotential height values for the 300-, 500-, and 850-hPa levels as well as the surface chart. A deep 500-hPa trough can be seen from Minnesota to Oklahoma with multiple areas of low pressure stretching along a frontal zone down the spine of the Appalachian Mountains. The SCEPT event on this date occurred in northeastern North Carolina ahead of the advancing cold front similarly to the synoptic type described by Maddox et al. (1979).

Most (78%) of the ULT cases occurred during the warm season, from May to August (Table 3), when more moisture and thermal instability were available for convective development. Thus, the strongest synoptic dynamics were not required for a ULT SCEPT event to develop during the warm season. With weaker synoptic dynamics, the events occurred closer to a moisture source, such as the Gulf of Mexico or Atlantic Ocean (Fig. 9). Most (66%) of the events developed south of 37°N. Little information was derived from the time of day with events scattered throughout the 24-h period (Fig. 5).

Most (89%) of the ULT SCEPT events occurred under weak to no positive differential vorticity advection, similar to past research (Moore et al. 2003). A 500-hPa long-wave trough was always evident on the synoptic charts, averaging 820 km (510 mi; Table 7) to the west along a line of equal latitude. Two-thirds of the time the trough was either slightly positively or neutrally tilted and was also strengthening at the time of the ULT SCEPT event.

Pertaining to the 300-hPa level, 89% of the events occurred when a trough was located to the west, averaging 915 km (565 miles; Table 7), along a line of equal latitude. Much like at 500 hPa, the trough was slightly positively or neutrally tilted and was also strengthening in two-thirds of the events.

The ULT cases only had a closed 850-hPa low in 22% of the events. However, the region of lowest height at 850 hPa was consistently to the northwest at an average distance of 760 km (470 mi; Fig. 10a). This placed the ULT SCEPT events under WAA during 70% of the situations, increasing instability and lift. Much like the CULL cases the 850-hPa jet was oriented south-southwest to north-northeast, traversing directly over the training site in most events (89%). The tilt of the 850-hPa trough was positive in a majority of the ULT SCEPT events (67%), but no relationship was determined with regard to whether the trough was strengthening or weakening.

At the surface, the cold front was a comparable distance away from the ULT SCEPT event as with the CULL case, averaging 460 km (285 mi; Table 7) to the west along a line of equal latitude. The warm front was less defined with SCEPT events occurring well north and well south of the boundary. On average, the warm front location was 190 km (120 mi; Table 7) to the north of the ULT SCEPT event along a line of equal longitude. The surface low averaged 530 km (330 mi; Fig. 10b) to the northwest. This usually placed the ULT SCEPT events in the warm sector, or along the warm front, of a midlatitude cyclone.

The greatest surface moisture convergence was clustered within 200 km (125 mi) and had little linear relationship (r = −0.07; Table 8). On average, the greatest surface moisture convergence was 65 km (40 mi) southeast of the ULT SCEPT event location. With slightly reduced synoptic dynamics, higher precipitable water values were needed to compensate. Precipitable water averaged 37.1 mm (1.46 in.; Table 7). Again, these values were slightly less than what recent research has found for warm season events (Schultz 1989; Roebber and Eise 2001) but exactly the same as was found in Maddox et al. (1979) for synoptically forced systems. The precipitable water tongue extended 710 km (440 mi) to the northeast of the training events (Fig. 11b), advecting from the south-southwest.

UVV values at 700 hPa were less than in the CULL events, averaging −0.37 Pa s−1. The reduction in UVVs was likely due to the slightly weaker dynamics compared to the CULL cases. A strong linear west-southwest to east-northeast relationship was found (r = 0.73, p < 0.05) in the location of the maximum UVV at 700 hPa nearest the ULT SCEPT event. Again, just like in the CULL events, the UVV maximum was found to be along the flow at 700 hPa intersecting with the SCEPT event location. The maximum UVVs averaged 180 km (110 mi) southwest of the ULT event.

In summary, ULT SCEPT events were defined by a 500-hPa trough without a closed low. Because of the lack of events meeting the ULT requirements, only 700-hPa maximum UVVs had a strong linear patter from west-southwest to east-northeast (r = 0.73, p < 0.05), indicating maximum UVVs at 700 hPa occurred along the flow at 700 hPa, which intersected with the ULT SCEPT event. Most of the events had an 850-hPa jet, helping to advect moisture from the south-southwest, resulting in precipitable water values averaging 37.1 mm (1.46 in). UVVs were weaker than in the CULL cases, but warm-air advection was still occurring in a majority of the events. The ULT SCEPT events occurred in the warm sector of a midlatitude cyclone or along the warm front with the cold front 460 km (285 mi) to the west. Strong troughs in the upper levels at 300 and 500 hPa were positively or negatively tilted and were also strengthening at the time of the ULT SCEPT event two-thirds of the time.

c. 850-hPa trough–low

Unlike the first two cases where long-wave 500-hPa troughs were present, the 850-hPa trough–low (850TL) case has fewer upper-level influences (above 700 hPa) on the environment of the SCEPT event (Table 9). Typically, a short-wave trough at 500 hPa helped to influence the event, but never was a long-wave trough present. However, a moderate 850-hPa low or trough was always present. Figure 12 includes geopotential height fields for 500, 300, and 850 hPa, and a surface chart for an 850TL event on 12 July 2004. A short-wave trough at 500 hPa can be seen in the northeast with a more substantial 850-hPa low centered in southern Ontario. The surface low pressure is indicated over Lake Michigan with a warm front or stationary boundary through central Pennsylvania. The SCEPT event occurred in south-central Pennsylvania along this boundary, which is close to, but not exactly like, the frontal type described by Maddox et al. (1979).

850TL SCEPT events occurred primarily (90%) during the warm season (May–August; Table 3) when abundant moisture was present. Like the CULL cases, many of the events (70%) occurred in the Midwest (Fig. 13). During the warm season, the Midwest is more susceptible to incursions of frontal boundaries and influences from the upper levels than is the South due to a building ridge in the Southeast; therefore, a short-wave trough at 500 hPa could influence an event in the Midwest, but not in the Southeast. This might indicate why most of the events occurred farther north even though more thermal instability and moisture were located closer to the Gulf of Mexico. The maximum number of events occurred during the 0800–1600 UTC time frame, the time of least thermal instability (Fig. 5).

The primary difference between the 850TL cases and the previous two cases was weaker upper-level synoptic flow at 500 hPa and above with no long-wave troughs present. Little information was obtained from these levels as a 500-hPa trough was only present in 40% of the events, which was further reduced to 20% at the 300-hPa level.

On average, the location of the lowest height at the 850-hPa level was 565 km (350 mi) to the west of the 850TL SCEPT events (Fig. 14a). A moderately strong linear pattern (r = 0.50, p < 0.10; Table 10) was found in the location of the 850-hPa low pressure system.

WAA occurred in a majority of the cases (60%), while little or no advection was present in 40% of the events. Thermal instability was likely a greater contributor to lift rather than the dynamics in the region. No relationship could be identified in regard to whether the trough was positively or negatively tilted. However, the 850-hPa trough was either strengthening or remaining steady 90% of the time.

At the surface, cold and warm fronts were commonly associated with 850TL cases. The cold front averaged 300 km (185 mi; Table 11) west of the 850TL SCEPT event locations along a line of equal latitude, while the warm front averaged 85 km (55 mi; Table 11) to the north along a line of equal longitude. This positioned the event either along the warm front or in the warm sector of a midlatitude cyclone. Without the strong 500-hPa trough, 850TL SCEPT events do not specifically match either of the Maddox et al. (1979) heavy-rain types. The surface low was oriented from the west to the northeast of the 850TL SCEPT events in a strong linear pattern (r = 0.80, p < 0.05), averaging 560 km (350 mi) to the west-northwest of the 850TL SCEPT events (Fig. 14b).

The same general trend with the CULL and ULT cases continued with the 850TL cases concerning surface moisture convergence. The greatest moisture convergence occurred along the flow of the 850-hPa jet either upstream, over, or downstream of the 850TL SCEPT event locations 70% of the time (Fig. 15a). This again disagrees with past research (Junker et al. 1999; Moore et al. 2003). A linear pattern (r = −0.31, Table 10) was found; however, it was not significant (p < 0.10). This was the only variable that did not have a significant linear pattern in the 850TL cases.

Precipitable water was considerably higher during an 850TL SCEPT event, averaging 42.2 mm (1.66 in.; Table 11), nearing the value found in past research for warm season events (Schultz 1989; Roebber and Eise 2001). With most of the events occurring during the warm season, moisture was readily available, overcoming the lack of a strongly forced synoptic environment. However, there were still similarities to the previous two cases with the apex of the moisture tongue located on average 290 km (180 mi) to the northeast of the 850TL SCEPT events (Fig. 15b) with a moderately strong linear pattern from southwest to northeast. In addition, moisture was advected on average from the south-southwest.

With the lack of a strongly forced synoptic environment, UVVs averaged only −0.39 Pa s−1. However, a moderately strong linear pattern (r = −0.43, p < 0.05; Table 10), from west-northwest to east-southeast, was found in proximity of the maximum UVVs compared to the 850TL SCEPT event locations (Fig. 15c).

In summary, an 850TL case was defined by weak upper-level flow and an 850-hPa trough or low present, on average, 565 km (350 mi) to the west of the 850TL SCEPT event location. All variables examined in this study had a moderately strong linear pattern from southwest–west-northwest to northeast–east-southeast of the 850TL SCEPT events, except for surface moisture convergence. A linear pattern was still found with surface moisture convergence, but it was not significant (p < 0.10). Surface moisture convergence was greatest along the flow of the 850-hPa jet. The maximum location of the UVVs at 700 hPa occurred along the same latitude as the 850TL event. Precipitable water averaged 42.2 mm (1.66 in.), near values found in past research for heavy-rain events in the warm season. The moisture was advected from the south-southwest with a maximum extent of the moisture tongue 290 km (180 mi) north-northeast of the events. Upper-level forcing was weak and no relationships could be found with 500- and 300-hPa flows.

6. Conclusions

Forecasting when synoptically forced extreme precipitation training convection will occur is important because of its enhanced risk of flash flooding. Little research has been done on indicating the locations in which training convection occurs in a synoptically forced environment (Moore et al. 2003; Junker et al. 1999) even though a statistical increase in 500-hPa cyclones has been found in the last 50 yr (Key and Chan 1999). This paper established three separate cases in which extreme precipitation occurred because of training convection during a synoptically forced event (SCEPT). The three cases identified were the closed upper-level low (CULL), the upper-level trough (ULT), and the 850-hPa trough–low (850TL). The first two cases were defined by a long-wave trough at 500 hPa; however, a CULL event was the only type in which a closed upper-level low was present at 500 hPa or above. Long-wave troughs were not present during 850TL events.

Many of the variables investigated in this study confirmed past findings regarding heavy rainfall. High precipitable water content was required in every event type ranging from 36.2 mm (1.42 in.) for CULL events to 42.2 mm (1.66 in.) for 850TL events. This is consistent with past research, particularly Maddox et al. (1979). Midlatitude low pressure systems and upper-level lows were often positioned to the west or northwest of the SCEPT event locations, with events generally occurring near frontal boundaries, in particular the warm front. Additionally, strong surface moisture convergence and upward vertical velocities at 700 hPa were found. Little differential vorticity advection influenced the extreme rainfall events in this study.

All of the variables in the previous paragraph are consistent from past research described in section 2 and would only help to locate a broad area in which heavy rainfall will occur. However, heavy rainfall does not necessarily mean that training convection or extreme rainfall is imminent. Thus, additional research was done to help decipher when, and where, a training extreme precipitation event was possible.

Maddox et al. (1979) found that during a heavy rainfall event the 500-hPa trough was nearly stationary. In this research project it was found that often the 500-hPa trough for CULL and ULT events (not 850TL events due to the lack of a long-wave trough at 500 hPa) was slightly positively to neutrally tilted. Often when this occurs, the trough is in a transition mode, resulting in little movement, or nearly stationary. However, the transition mode does not mean the trough was weakening; rather it was found to be strengthening or deepening at the time of the SCEPT event. This again confirms the heavy rainfall research conducted in the past, but this is often vital to the existence of a training convection situation in a synoptically forced environment. Additionally, SCEPT events were found to develop on the right side of a 300-hPa jet streak with no preference toward a quadrant determined.

As stated earlier, the existence of strong surface moisture convergence and 700-hPa upward vertical velocities in the vicinity of heavy rainfall is well known. However, the positioning of these variables in relation to the location of training convection had not yet been determined. It was found that the local maximum of surface moisture convergence was generally within 230 km of the SCEPT event location either upstream, downstream, or over the area of training parallel to the 850-hPa flow in a linear pattern. Additionally, maximums in 700-hPa upward vertical velocities were generally within 270 km of the SCEPT event parallel to the flow again upstream, downstream, or over the area of training at 700 hPa from west-southwest to east-northeast again in a linear pattern. The linear patterns are important because they allow the forecaster to outlook for an area in which a SCEPT event is possible around 700-hPa upward vertical velocities and local maximums of surface moisture convergence. Looking in these regions will help locate an area of training convection within a heavy-rain event, which would become extreme should precipitable water values be in accordance with our above-described findings, resulting in a dangerous flash flooding situation.

Future research is needed to determine why many of the SCEPT events occurred during the period of least thermal instability in the early morning hours. It is possible that a weakening low-level jet during the morning hours leads to a reduction in the forward speed of the convective system and, thus, an increase in the potential rainfall over a specific point, as described by Doswell et al. (1996). A longer research period would be helpful in the future to increase the significance of the linearity of many of the variables in this study. Additionally, a finer look into the magnitudes of surface moisture convergence and upward vertical velocities could reduce the potential for false positives.

Acknowledgments

The authors thank Dr. P. Grady Dixon, Dr. Michael E. Brown, Mrs. Shaina Niehans, and two anonymous reviewers for their help in editing the paper. Additionally, this paper would never have been possible without the help of Mississippi State University’s Department of Geosciences.

REFERENCES

  • Beebe, R. G., , and Bates F. C. , 1955: A mechanism for assisting in the release of convective instability. Mon. Wea. Rev., 83 , 110.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bradley, A. A., , and Smith J. A. , 1994: The hydrometeorological environment of extreme rainstorms in the southern plains of the United States. J. Appl. Meteor., 33 , 14181431.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Branick, M. L., , Vitale F. , , Lai C. C. , , and Bosart L. F. , 1988: The synoptic and subsynoptic structure of a long-lived severe convective system. Mon. Wea. Rev., 116 , 13351370.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , and Stensrud D. J. , 2000: Climatology of heavy rain events in the United States from hourly precipitation observations. Mon. Wea. Rev., 128 , 11941201.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Byers, H. R., , and Braham R. R. Jr., 1949: The Thunderstorm: Report of the Thunderstorm Project. U.S. Government Printing Office, 300 pp.

    • Search Google Scholar
    • Export Citation
  • Chumchean, S., , Seed A. , , and Sharma A. , 2004: Application of scaling in radar reflectivity for correcting range-dependent bias in climatological radar rainfall estimates. J. Atmos. Oceanic Technol., 21 , 15451555.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Collier, C. G., 1996: Applications of Weather Radar System: A Guide to Uses of Radar in Meteorology and Hydrology. John Wiley and Sons, 294 pp.

    • Search Google Scholar
    • Export Citation
  • Colman, B. R., 1984: Deep convection occurring above a stable planetary boundary layer. Ph.D. thesis, Massachusetts Institute of Technology, 243 pp.

  • Colorado State University, cited. 2009: NWS DIFAX weather map archive. [Available online at http://archive.atmos.colostate.edu.

  • Crook, N. A., , Carbone R. E. , , Moncrieff M. W. , , and Conway J. W. , 1990: The generation and propagation of a nocturnal squall line. Part II: Numerical simulations. Mon. Wea. Rev., 118 , 5066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. S., 2001: Flash flood forecast and detection methods. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 481–525.

    • Search Google Scholar
    • Export Citation
  • Doswell C. A. III, , , Brooks H. E. , , and Maddox R. A. , 1996: Flash flood forecasting: An ingredients-based methodology. Wea. Forecasting, 11 , 560581.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fabry, F., , Bellon A. , , Duncan M. R. , , and Austin G. L. , 1994: High resolution rainfall measurements by radar for very small basins: The sampling problem re-examined. J. Hydrol., 161 , 428.

    • Search Google Scholar
    • Export Citation
  • Fulton, R. A., , Breidenback J. P. , , Seo D-J. , , and Miller D. A. , 1998: The WSR-88D rainfall algorithm. Wea. Forecasting, 13 , 377395.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Heideman, K. F., , and Fritsch J. M. , 1988: Forcing mechanisms and other characteristics of significant summertime precipitation. Wea. Forecasting, 3 , 115130.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hildebrand, P. H., 1978: Iterative correction for attenuation of 5 cm radar in rain. J. Appl. Meteor., 17 , 508514.

  • Junker, N. W., , Schneider R. S. , , and Fauver S. L. , 1999: A study of heavy rainfall events during the Great Midwest Flood of 1993. Wea. Forecasting, 14 , 701712.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Key, J. R., , and Chan A. C. K. , 1999: Multidecadal global and regional trends in 1000 mb and 500 mb cyclone frequencies. Geophys. Res. Lett., 26 , 20532056.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Konrad C. E. II, , 2001: The most extreme precipitation events over the eastern United States from 1950 to 1996: Considerations of scale. J. Hydrometeor., 2 , 309325.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maddox, R. A., , and Doswell C. A. III, 1982: An examination of jet stream configurations, 500 mb vorticity advection and low-level thermal advection patterns during extended periods of intense convection. Mon. Wea. Rev., 110 , 184197.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maddox, R. A., , Chappel C. F. , , and Hoxit L. R. , 1979: Synoptic and meso-α scale aspects of flash flood events. Bull. Amer. Meteor. Soc., 60 , 115123.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, R. C., 1967: Notes on analysis and severe storm forecasting procedures of the Military Weather Warning Center. AWS Tech. Rep. 200, 94 pp. [Available from Headquarters Air Weather Service, Scott AFB, IL 62225].

    • Search Google Scholar
    • Export Citation
  • Moore, J. T., , Glass F. H. , , Graves C. E. , , Rochette S. M. , , and Singer M. J. , 2003: The environment of warm-season elevated thunderstorms associated with heavy rainfall over the central United States. Wea. Forecasting, 18 , 861878.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1950: The structure and mechanism of the prefrontal squall line. J. Meteor., 7 , 210222.

  • Newton, C. W., 1963: Dynamics of severe convective storms. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 33–58.

  • Niziol, T. A., , Snyder W. R. , , and Waldstreicher J. S. , 1995: Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow. Wea. Forecasting, 10 , 6177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NWS, cited. 2007: WSR-88D rainfall estimation. HL publications and papers. [Available online at http://www.nws.noaa.gov/oh/hrl/papers/papers.htm#wsr88d.

    • Search Google Scholar
    • Export Citation
  • NWS, cited. 2009: National Weather Service glossary. [Available online at http://forecast.weather.gov/glossary.php?word=precipitable+water.

    • Search Google Scholar
    • Export Citation
  • Rhea, O. J., 1966: A study of thunderstorm formation along dry lines. J. Appl. Meteor., 5 , 5863.

  • Riehl, H., and Coauthors, 1952: Forecasting in the Middle Latitudes. Meteor. Monogr., No. 5, Amer. Meteor. Soc., 80 pp.

  • Roebber, P. J., , and Eise J. , 2001: The 21 June 1997 flood: Storm-scale simulations and implications for operational forecasting. Wea. Forecasting, 16 , 197218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rose, S. F., , Hobbs P. V. , , Locatelli J. D. , , and Stoelinga M. T. , 2004: A 10-yr climatology relating the locations of reported tornadoes to the quadrants of upper-level jet streaks. Wea. Forecasting, 19 , 301309.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , Atlas D. , , and Short D. A. , 1990: The estimation of rainfall by area integrals. Part 2. The height–area rain threshold (HART) method. J. Geophys. Res., 95 , 21612176.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ross, B. B., 1987: The role of low-level convergence and latent heating in a simulation of observed squall line formation. Mon. Wea. Rev., 115 , 22982321.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rotunno, R., , Klemp J. B. , , and Weisman M. L. , 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463485.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schultz, P., 1989: Relationships of several stability indices to convective weather events in northeast Colorado. Wea. Forecasting, 4 , 7380.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , and Johnson R. H. , 2005: Organization and environmental properties of extreme-rain-producing mesoscale convective systems. Mon. Wea. Rev., 133 , 961976.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , and Johnson R. H. , 2006: Characteristics of U.S. extreme rain events during 1999–2003. Wea. Forecasting, 21 , 6985.

  • Seo, D-J., 1998: Real-time estimation of rainfall fields using radar rainfall and rain gauge data. J. Hydrol., 208 , 2536.

  • Short, D. A., , Kozu T. , , and Nakamura K. , 1990: Rainrate and raindrop size distribution observations in Darwin, Australia. Proc. URSI-F Open Symp. on Regional Factors in Predicting Radiowave Attenuation due to Rain, Rio de Janeiro, Brazil, Rutherford Appleton Laboratory and Empresa Brasileira de Telecomunicaoes, 35–40.

    • Search Google Scholar
    • Export Citation
  • Smith, J. A., , and Krajewski W. F. , 1991: Estimation of the mean field bias of radar rainfall estimates. J. Appl. Meteor., 30 , 397412.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, E. N., , Hollinger S. E. , , and Kunkel K. E. , 2005: Use of real-time multisensor data to assess the relationship of normalized corn yield with monthly rainfall and heat stress across the central United States. J. Appl. Meteor., 44 , 16671676.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • White, G. H., 1982: An observational study of the Northern Hemisphere extratropical summertime global circulation. J. Atmos. Sci., 39 , 2440.

    • Crossref
    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

The study region is shaded.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 2.
Fig. 2.

Multisensor precipitation estimates from radar rainfall data for the data range 1–31 mm h−1. (top) Histogram of all hourly rainfall rates for all good data in the study area. Any data points with hourly rainfall rates greater than 31 mm h−1 have been removed. (bottom) Cumulative total of all hourly rainfall rates, showing that the 99th percentile occurs at 16 mm h−1.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 3.
Fig. 3.

Example CULL SCEPT event from 2 Apr 2005, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) the surface analysis provided by the Colorado State University DIFAX archive.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 4.
Fig. 4.

CULL event locations.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 5.
Fig. 5.

Median times during which training occurred for all CULL, ULT, and 850TL events.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 6.
Fig. 6.

CULL events indicating the (a) location of the 500-hPa low, (b) location of the 300-hPa low, (c) location of the 850-hPa low, and (d) location of the surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 7.
Fig. 7.

CULL events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 8.
Fig. 8.

Example ULT SCEPT event from 2 May 2004, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) the surface analysis provided by the Colorado State University DIFAX archive.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 9.
Fig. 9.

ULT event locations.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 10.
Fig. 10.

ULT events indicating the (a) location of the 850-hPa low and (b) surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 11.
Fig. 11.

ULT events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) the location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 12.
Fig. 12.

Example 850TL SCEPT event from 12 Jul 2004, with event location marked with an x. Geopotential height from the NARR dataset for (a) 300, (b) 500, and (c) 850 hPa, and (d) at the surface analysis provided by the Colorado State University DIFAX archive.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 13.
Fig. 13.

850TL event locations.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 14.
Fig. 14.

850TL events indicating the (a) location of the 850-hPa low and (b) surface low in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Fig. 15.
Fig. 15.

850TL events indicating the (a) location of the maximum surface moisture convergence, (b) greatest extent of the precipitable water moisture tongue, and (c) location of the maximum UVV in relation to the location where the SCEPT event occurred. The squares indicate the average locations for the respective variables.

Citation: Weather and Forecasting 25, 2; 10.1175/2009WAF2222275.1

Table 1.

Total synoptically forced, tropical, MCS (nonsynoptically forced), and airmass extreme rainfall events of greater than 17 mm h−1 for three consecutive hours by year.

Table 1.
Table 2.

CULL event dates, longitudes, latitudes, time of day, and total rainfall amounts [mm (in.)].

Table 2.
Table 3.

Percentages of occurrence during 3-month periods for each of the three cases. The majority of the events for the CULL case occurred during the cool season, and in the warm season for the ULT and 850TL cases.

Table 3.
Table 4.

Standard distance, Pearson’s R value, and correlation significance for CULL events.

Table 4.
Table 5.

CULL event statistics for average precipitable water (mm), cold front distance west (km), and warm front distance north (km).

Table 5.
Table 6.

ULT event dates, latitudes, longitudes, times of day, and total rainfall amounts [mm (in.)].

Table 6.
Table 7.

ULT event statistics for average precipitable water (mm), cold front distance west (km), warm front distance north (km), and 300- and 500-hPa trough distance west (km).

Table 7.
Table 8.

Standard distance, Pearson’s R value, and correlation significance for ULT events.

Table 8.
Table 9.

850TL event dates, latitudes, longitudes, times of day, and total rainfall amounts [mm (in.)].

Table 9.
Table 10.

Standard distance, Pearson’s R value, and significance for 850TL events.

Table 10.
Table 11.

850TL event statistics for average precipitable water (mm), cold front distance (km), and warm front distance (km).

Table 11.
Save