• Arritt, R. W., , Rink T. D. , , Segal M. , , Todey D. P. , , Clark C. A. , , Mitchell M. J. , , and Labas K. M. , 1997: The Great Plains low-level jet during the warm season of 1993. Mon. Wea. Rev., 125 , 21762192.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, G. D., , and Bosart L. F. , 1989: A 15-year climatology of Northern Hemisphere 500 mb closed cyclone and anticyclone centers. Mon. Wea. Rev., 117 , 21422164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bierly, G. D., , and Winkler J. A. , 2001: A composite analysis of airstreams within cold-season Colorado cyclones. Wea. Forecasting, 16 , 5780.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bjornsson, H., , Mysak L. A. , , and Brown R. D. , 1995: On the interannual variability of precipitation and runoff in the Mackenzie drainage basin. Climate Dyn., 12 , 6776.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bonner, W. D., 1968: Climatology of the low-level jet. Mon. Wea. Rev., 96 , 833850.

  • Brubaker, K. L., , Dirmeyer P. A. , , Sudradjat A. , , Levy B. S. , , and Bernal F. , 2001: A 36-yr climatological description of the evaporative sources of warm-season precipitation in the Mississippi River basin. J. Hydrometeor., 2 , 537557.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chung, Y., , Hage K. D. , , and Reinelt E. R. , 1976: On lee cyclogenesis and airflow in the Canadian Rocky Mountains and the East Asian Mountains. Mon. Wea. Rev., 104 , 879891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., , and Brubaker K. L. , 1999: Contrasting evaporative moisture sources during the drought of 1988 and the flood of 1993. J. Geophys. Res., 104 , 1938319397.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., , 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
  • Draxler, R. R., 2003: Evaluation of an ensemble dispersion calculation. J. Appl. Meteor., 42 , 308317.

  • Draxler, R. R., , and Hess G. D. , 1997: Description of the HYSPLIT_4 modeling system. NOAA Tech. Memo. ERL ARL-224, 24 pp.

  • Draxler, R. R., , and Hess G. D. , 1998: An overview of the HYSPLIT_4 modelling system for trajectories, dispersion, and deposition. Aust. Meteor. Mag., 47 , 295308.

    • Search Google Scholar
    • Export Citation
  • Fink, A. H., , and Knippertz P. , 2003: An extreme precipitation event in southern Morocco in spring 2002 and some hydrological implications. Weather, 58 , 377387.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fritsch, J. M., , Kane R. J. , , and Chelius C. R. , 1986: The contribution of mesoscale convective weather systems to the warm-season precipitation in the United States. J. Appl. Meteor., 25 , 13331345.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Helfand, H. M., , and Schubert S. D. , 1995: Climatology of the simulated Great Plains low-level jet and its contribution to the continental moisture budget of the United States. J. Climate, 8 , 784806.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , Yao Y. , , Yarosh E. S. , , Janowiak J. E. , , and Mo K. C. , 1997: Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States. J. Climate, 10 , 481507.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hoke, J. E., , Phillips N. A. , , Dimego G. J. , , Tuccillo J. J. , , and Sela J. G. , 1989: The regional analysis and forecast system of the national meteorological center. Wea. Forecasting, 4 , 323334.

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

  • Kanamitsu, M., 1989: Description of the NMC global data assimilation and forecast system. Wea. Forecasting, 4 , 335342.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Keim, B. D., 1998: Record precipitation totals from the coastal New England rainstorm of 20–21 October 1996. Bull. Amer. Meteor. Soc., 79 , 10611068.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lackmann, G. M., , Gyakum J. R. , , and Benoit R. , 1998: Moisture transport diagnosis of a wintertime precipitation event in the Mackenzie River basin. Mon. Wea. Rev., 126 , 668692.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lanicci, J. M., , and Warner T. T. , 1991: A synoptic climatology of the elevated mixed-layer inversion over the southern Great Plains in spring. Part I: Structure, dynamics, and seasonal evolution. Wea. Forecasting, 6 , 181197.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y., , Chiao S. , , Wang T. , , Kaplan M. , , Michael L. , , and Weglarz R. P. , 2001: Some common ingredients for heavy orographic rainfall. Wea. Forecasting, 16 , 633660.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, J., , and Stewart R. E. , 2003: Water vapor fluxes over the Saskatchewan River basin. J. Hydrometeor., 4 , 944959.

  • Louie, P. Y. T., , Hogg W. D. , , MacKay M. D. , , Zhang X. , , and Hopkinson R. F. , 2002: The water balance climatology of the Mackenzie River basin with reference to the 1994/1995 water year. Atmos.–Ocean, 49 , 159180.

    • Search Google Scholar
    • Export Citation
  • Mitchell, M. J., , Arritt R. W. , , and Labas K. , 1995: A climatology of the warm season Great Plains low-level jet using wind profiler observations. Wea. Forecasting, 10 , 576591.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Novak, M. J., , Bosart L. F. , , Keyser D. , , Wasula T. A. , , and LaPenta K. D. , 2002: Climatology of warm season 500 hPa cutoff cyclones and a case study diagnosis of 14–17 July 2000. Preprints, 19th Conf. on Weather Analysis and Forecasting, San Antonio, TX, Amer. Meteor. Soc., 68–71.

  • Pan, Z., , Segal M. , , and Arritt R. W. , 2004: Role of topography in forcing low-level jets in the central United States during the 1993 flood-altered terrain simulations. Mon. Wea. Rev., 132 , 396403.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulos, G. S., , Wesley D. A. , , Snook J. S. , , and Meyers M. P. , 2002: A Rocky Mountain storm. Part I: The blizzard—Kinematic evolution and the potential for high-resolution numerical forecasting of snowfall. Wea. Forecasting, 17 , 955970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Proctor, B. A., , Wang M. , , Strong G. S. , , and Smith C. D. , 1999: Atmospheric moisture budgets for MAGS. Proc. Fifth Scientific Workshop for the Mackenzie GEWEX Study (MAGS), Edmonton, AB, Canada, Environment Canada and the Natural Sciences and Engineering Research Council of Canada, 117–124.

  • Randel, D. L., , Greenwald T. J. , , Vonder Haar T. H. , , Stephens G. L. , , Ringerud M. A. , , and Combs C. L. , 1996: A new global water vapor dataset. Bull. Amer. Meteor. Soc., 77 , 12331246.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reiff, J., , Forbes G. S. , , Spieksma F. T. M. , , and Reynders J. J. , 1986: African dust reaching northwestern Europe: A case study to verify trajectory calculations. J. Climate Appl. Meteor., 25 , 15431567.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reuter, G. W., , and Nguyen C. D. , 1993: Organization of cloud and precipitation in an Alberta storm. Atmos. Res., 30 , 127141.

  • Rolph, G. D., , and Draxler R. R. , 1990: Sensitivity of three-dimensional trajectories to the spatial and temporal densities of the wind field. J. Appl. Meteor., 29 , 10431054.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Salati, E., , Dall’Olio A. , , Matsui E. , , and Gat J. R. , 1979: Recycling of water in the Amazon basin: An isotopic study. Water Resour. Res., 15 , 12501258.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schubert, S. D., , Helfand H. M. , , Wu C. , , and Min W. , 1998: Subseasonal variations in warm-season moisture transport and precipitation over the central and eastern United States. J. Climate, 11 , 25302555.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smirnov, V. V., , and Moore G. W. K. , 2001: Short-term and seasonal variability of the atmospheric water vapor transport through the Mackenzie River basin. J. Hydrometeor., 2 , 441452.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stewart, R. E., and Coauthors, 1998: The Mackenzie GEWEX Study: The water and energy cycles of a major North American river basin. Bull. Amer. Meteor. Soc., 79 , 26652684.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stohl, A., 1998: Computation, accuracy and applications of trajectories—A review and bibliography. Atmos. Environ., 32 , 947966.

  • Strong, G. S., , Proctor B. , , Wang M. , , Soulis E. D. , , Smith C. D. , , Seglenieks F. , , and Snelgrove K. , 2002: Closing the Mackenzie drainage basin water budget, water years 1994/95 to 1996/97. Atmos.–Ocean, 40 , 113124.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Szeto, K. K., 2002: Moisture recycling over the Mackenzie Basin. Atmos.–Ocean, 40 , 181197.

  • Ulbrich, U., , Brucher T. , , Fink A. H. , , Leckebusch G. C. , , Kruger A. , , and Pinto J. G. , 2003: The central European floods of August 2002. Part II: Synoptic causes and considerations with respect to climate change. Weather, 58 , 434442.

    • Search Google Scholar
    • Export Citation
  • Verschuren, J. P., , and Wojtiw L. , 1980: Estimate of the maximum probable precipitation for Alberta river basins. Alberta Environment, Hydrology Branch, RMD-80/1, Edmonton, Alberta, 307 pp.

  • Walters, C. K., , and Winkler J. A. , 2001: Airflow configurations of warm season southerly low-level wind maxima in the Great Plains. Part I: Spatial and temporal characteristics and relationship to convection. Wea. Forecasting, 16 , 513530.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Whittaker, L. M., , and Horn L. H. , 1981: Geographical and seasonal distribution of North American cyclogenesis, 1958–1977. Mon. Wea. Rev., 109 , 23122322.

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

    A map of the Mackenzie River basin (MRB) and Saskatchewan River basin. Subbasins over the southern MRB are delineated by gray lines. SWH represents the Swan Hills, and SP represents the upper-air sounding site at Stony Plain. Points labeled A1 to A3 and P1 to P3 represent the locations of streamflow gauges discussed in the text.

  • View in gallery

    Forty-eight-hour accumulated rainfall (mm) for three extreme rainfall events over Alberta: (a) 22–23 Jun 1993, (b) 18–19 Jun 1996, and (c) 28–29 Jul 2001. The two polygons enclosed by the solid dark lines over northern Alberta approximate the boundaries of the Peace River and Athabasca River basins (see Fig. 1 for details).

  • View in gallery

    Streamflow measurements (m3 s−1) at selected locations in the Peace River and Athabasca River basins: (a) 22–23 Jun 1993, (b) 18–19 Jun 1996, and (c) 28–29 Jul 2001. Refer to Fig. 1 for the locations of streamflow measurement sites A1–A3 and P1–P3. Vertical tick marks indicate time of maximum streamflow associated with each event.

  • View in gallery

    (left) The 500-hPa and (right) mean sea level pressure maps for the 22–23 Jun 1993, 18–19 Jun 1996, and 28–29 Jul 2001 extreme rainfall events. Contour interval for the 500-hPa heights is 60 gpm, and for the mean sea level pressure is 4 hPa. (Images provided by the NOAA–CIRES Climate Diagnostics Center Web site at www.cdc.noaa.gov)

  • View in gallery

    (left) Continuous trajectories from the Gulf of Mexico to the southern MRB for the three extreme rainfall events. Solid triangles are shown every 24 h. (right) The corresponding pressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

  • View in gallery

    (top) Backward ensemble trajectories calculated from the southern MRB for the 22–23 Jun 1993 extreme rainfall event. Trajectories labeled P and G originate from the North Pacific Ocean and the Gulf of Mexico, respectively. (bottom) The evolution in height of the trajectories (hPa).

  • View in gallery

    (left) Stepwise trajectory for the Jun 1996 extreme rainfall event. Solid triangles are shown every 12 h. (right) The correspondingpressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

  • View in gallery

    (left) Stepwise trajectory for the Jul 2001 extreme rainfall event. Solid triangles are shown every 12 h. (right) The correspondingpressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

  • View in gallery

    Same as for Fig. 6, except for backward ensemble trajectories calculated from the southern MRB for the 28–29 Jul 2001 extreme rainfall event.

  • View in gallery

    Schematic of the proposed conceptual model that describes the transport of low-level moisture from the Gulf of Mexico to northern Alberta. Here CL refers to the 500-hPa cutoff low, SL to the surface low, and GPLLJ to the Great Plains low-level jet. The shaded area denotes the region where mesoscale convective complexes occur. The star indicates the location of heavy rainfall. The dashed line and block arrows represent an idealized trajectory followed by the modified low-level Gulf moisture.

  • View in gallery

    The 850-hPa merdional wind anomalies (m s−1) calculated for (a) 18–22 Jun 1993, (b) 10–18 Jun 1996, and (c) 20–28 Jul 2001. The anomalies were calculated with respect to the 1948–96 means. Solid and dashed lines represent positive and negative anomalies, respectively. Anomalies larger than +1 m s−1 are shaded. (Images provided by the NOAA–CIRES Climate Diagnostics Center Web site at www.cdc.noaa.gov)

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 69 69 11
PDF Downloads 32 32 6

Transport of Atmospheric Moisture during Three Extreme Rainfall Events over the Mackenzie River Basin

View More View Less
  • 1 Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, Canada
© Get Permissions
Full access

Abstract

Lagrangian trajectories were computed for three extreme summer rainfall events (with rainfall exceeding 100 mm) over the southern Mackenzie River basin to test the hypothesis that the low-level moisture feeding these rainstorms can be traced back to the Gulf of Mexico. The three-dimensional trajectories were computed using the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT).

For all three events, parcel trajectories were identified that originated near the Gulf of Mexico and terminated over the southern Mackenzie River basin. Specifically, the transport of low-level moisture was found to occur along either quasi-continuous or stepwise trajectories. The time required to complete the journey varied between 6 and 10 days.

Closer examination of the data suggests that, for the three cases in question, the transport of modified Gulf of Mexico moisture to high latitudes was realized when the northward extension of the Great Plains low-level jet to the Dakotas occurred in synch with rapid cyclogenesis over Alberta, Canada. In this way, modified low-level moisture from the Gulf of Mexico arrived over the northern Great Plains at the same time as a strong southerly flow developed over the Dakotas and Saskatchewan, Canada, in advance of the deepening cutoff low over Alberta. This moist air was then transported northward over Saskatchewan and finally westward over the southern Mackenzie River basin, where strong ascent occurred.

Corresponding author address: Julian Charles Brimelow, Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta T6G 2E3, Canada. Email: brimelow@ualberta.ca

Abstract

Lagrangian trajectories were computed for three extreme summer rainfall events (with rainfall exceeding 100 mm) over the southern Mackenzie River basin to test the hypothesis that the low-level moisture feeding these rainstorms can be traced back to the Gulf of Mexico. The three-dimensional trajectories were computed using the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT).

For all three events, parcel trajectories were identified that originated near the Gulf of Mexico and terminated over the southern Mackenzie River basin. Specifically, the transport of low-level moisture was found to occur along either quasi-continuous or stepwise trajectories. The time required to complete the journey varied between 6 and 10 days.

Closer examination of the data suggests that, for the three cases in question, the transport of modified Gulf of Mexico moisture to high latitudes was realized when the northward extension of the Great Plains low-level jet to the Dakotas occurred in synch with rapid cyclogenesis over Alberta, Canada. In this way, modified low-level moisture from the Gulf of Mexico arrived over the northern Great Plains at the same time as a strong southerly flow developed over the Dakotas and Saskatchewan, Canada, in advance of the deepening cutoff low over Alberta. This moist air was then transported northward over Saskatchewan and finally westward over the southern Mackenzie River basin, where strong ascent occurred.

Corresponding author address: Julian Charles Brimelow, Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta T6G 2E3, Canada. Email: brimelow@ualberta.ca

1. Introduction

The Mackenzie River basin (MRB) is one of the world’s largest high-latitude river basins and covers an area of approximately 1.8 million km2 (Fig. 1). In this paper, we focus our attention on three extreme rainfall events (with rainfall exceeding 100 mm) that occurred over the southern MRB between 1993 and 2001: 22–23 June 1993, 18–19 June 1996, and 28–29 July 2001. The southern MRB encompasses the Peace River and Athabasca River basins, and covers an area of approximately 450 000 km2 (see Fig. 1). During the extreme rainfall events, portions of the southern MRB received between 100 and 150 mm of rain. These rainfall amounts represent about half of the mean summer rainfall over the southern MRB. Our objective was to determine whether the Gulf of Mexico (GOM) could act as a source of at least some of the moisture during extreme meso-α-scale summertime rainfall events over the southern MRB, and also to identify the mechanism(s) responsible for transporting the moisture-laden air.

Stewart et al. (1998) and Smirnov and Moore (2001) stressed the importance of understanding the atmospheric processes that transport water vapor into the MRB, because changes in the hydrological processes within the river basin can have important consequences for the regional climate. For example, Strong et al. (2002) found that the MRB acts as a moisture sink for most of the year, but during the summer months, can act as a source of moisture on account of enhanced evapotranspiration when soil moisture is high. Consequently, summer precipitation is an important part of the MRB’s hydrological cycle.

Lackmann et al. (1998) demonstrated that midlevel moisture from the Gulf of Alaska was crucial for wintertime precipitation in the MRB. Subsequent research by Smirnov and Moore (2001) found that during the autumn, winter, and spring months, extratropical cyclones were responsible for transporting midlevel moisture from the subtropical and midlatitude Pacific Ocean across the western boundary of the MRB. Liu and Stewart (2003) showed that the moisture-laden air entering the Saskatchewan River basin during the summer could originate from as far south as the GOM. The Saskatchewan River basin lies immediately to the southeast of the MRB and the two basins share a common boundary (see Fig. 1). Liu and Stewart (2003) also noted that the MRB receives moisture from the Saskatchewan River basin in the early summer. This finding is supported by Proctor et al. (1999), who found that the MRB loses moisture across its southeastern boundary in all months except June.

The origin of the low-level moisture for the three cases presented herein will be determined by computing three-dimensional trajectories using archived atmospheric fields from the National Centers for Environmental Prediction (NCEP) and the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT; Draxler and Hess 1998). Recently, Fink and Knippertz (2003) and Ulbrich et al. (2003) demonstrated the utility of employing trajectory models to identify remote sources of moisture during heavy rainfall events over northern Africa and Europe.

This paper provides the first quantitative investigation to determine whether the GOM can act as a potential source of moisture during summertime extreme rainfall events over the southern MRB. Our research will contribute to improving our understanding of the MRB’s hydrological cycle and supplement work already completed as part of the Mackenzie Global Energy and Water Cycle Experiment (GEWEX) Study (MAGS; Stewart et al. 1998).

2. Background

The mean annual precipitation over the southern MRB is approximately 500 mm (Louie et al. 2002), compared to 410 mm for the entire MRB (Bjornsson et al. 1995). Almost 50% of the annual precipitation over the southern MRB falls during the summer (Louie et al. 2002). A large proportion of the summer rainfall is from thunderstorms and is thus primarily dependent on surface or low-level moisture sources (Strong et al. 2002). Of particular importance, however, are meso-α-scale extreme rainfall events that are occasionally observed over the southern MRB in June and July (Verschuren and Wojtiw 1980). A single event producing 125 mm of rain can account for up to 25% of the annual rainfall over portions of the southern MRB. Extreme meso-α-scale rainfall events are almost exclusively associated with the passage of a 500-hPa cutoff low and rapid lee cyclogenesis over south-central Alberta (Reuter and Nguyen 1993). During the early summer months, the southern MRB is a climatologically preferred region for 500-hPa cutoff lows (Bell and Bosart 1989; Novak et al. 2002) and rapid lee cyclogenesis (Chung et al. 1976; Whittaker and Horn 1981). The rapidly intensifying surface low produces strong synoptic-scale lift and, when this lift is coupled with a moist air mass, widespread and heavy rainfall ensues. Case studies of heavy precipitation events associated with 500-hPa cutoff lows have been investigated by Reuter and Nguyen (1993), Keim (1998), and Poulos et al. (2002).

Verschuren and Wojtiw (1980) estimated the mean return period for a rainfall event producing 100 mm or more of rain over the foothills of central Alberta to be approximately 5 to 10 yr, increasing to 50 yr over far northern Alberta. The greatest frequency of extreme rainfall events occurs over the upper reaches of the Athabasca River basin in June. In those years when these events do occur, they have a significant impact on the annual water budget of the MRB on account of the large influx of moisture. This in turn has both immediate and short-term effects on the hydrological cycle through flooding and moisture recycling, respectively (Szeto 2002).

The rainfall rate of a rainstorm is proportional to the product of the water vapor mixing ratio of the air feeding the system and the vertical lift (e.g., Doswell et al. 1996). Thus, the likelihood of high rainfall rates decreases when the moisture content of the air is reduced. Operating under the assumption that the combination of strong synoptic-scale forcing or orographic lift is capable of producing significant and sustained lift over the threat area, the only factor precluding high rainfall rates is low values of the water vapor mixing ratio. It is well known that the total column water vapor content decreases steadily with increasing latitude (e.g., Randel et al. 1996). Thus, moisture from local sources, while important, is probably insufficient to account for the rainfall amounts observed during extreme rainfall events over the southern MRB, and there must be a remote source of very moist air to continuously replenish the water vapor as it is being converted to precipitation.

3. Data and methodology

a. Rainfall and streamflow data

Rainfall maps for the extreme rainfall events were created using 24-h accumulated rainfall data provided by Alberta Environment and the Meteorological Service of Canada. By synthesizing these rainfall data we were able to identify areas with high rainfall that would normally not be resolved by coarser rain gauge networks. The number of rainfall stations in Alberta that were used to create the maps for the 1993 event was 106, compared to 209 for the 1996 event and 258 for the 2001 event. Given that the rain for all three events lasted less than two days, 48-h accumulated rainfall maps were generated for each event. These maps were then used to quantify the intensity and spatial coverage of the rainfall.

To illustrate the effect of each rainfall event on the hydrology of the southern MRB, streamflow gauge data for selected monitoring sites (see Fig. 1) in the Athabasca and Peace River basins were obtained from Alberta Environment. The daily averaged data were used to create streamflow charts for the periods before, during, and following the event. In this way, the progression of the surge of water resulting from an extreme rainfall event could be tracked through the relevant river basin.

b. The HYSPLIT model and atmospheric input data

HYSPLIT is capable of computing forward or backward Lagrangian trajectories for an air parcel from any user-specified height and location (Draxler and Hess 1997, 1998). Hourly values of specified meteorological variables including the pressure, height, temperature, potential temperature, and relative humidity are provided along each trajectory.

HYSPLIT requires gridded 3D wind fields to compute the trajectories. Trajectories for the 1993 and 1996 events were computed using archived data from NCEP’s Regional Analysis and Forecast System (RAFS), which uses the Nested Grid Model (NGM) for the forecasts (Hoke et al. 1989). These data are archived every 2 h on a 33 by 28 polar stereographic grid (horizontal grid spacing of 180 km) with 10 data levels in the vertical. For the 2001 event, we used the Global Data Assimilation System (GDAS; Kanamitsu 1989), which uses the Medium Range Forecast model (MRF). The GDAS data are archived every 6 h on hemispheric 129 by 129 polar stereographic grids (equivalent to a horizontal grid spacing of 191 km) with 13 data levels in the vertical. This particular data archive was selected over that of the higher-resolution Eta Data Assimilation System (EDAS), because the Eta Model domain did not cover all of the study area.

The HYSPLIT model allows the user to simultaneously release parcels from all points within a user-specified matrix at a given time and height. Forward trajectories from the GOM to the southern MRB were identified as follows: parcels were released from locations on a 1° by 1° grid that encompassed the area between the NW and SE coordinates of 30°N, 100°W and 25°N, 85°W, respectively. These starting points did not necessarily have to be located on the RAFS or GDAS grid points, as HYSPLIT uses bilinear interpolation to calculate the meteorological variables at times and locations between the standard times and grid points available in the gridded datasets. The parcels were released at hourly intervals between 5 and 10 days prior to the heavy rainfall events (given the distances involved, this was deemed to be a reasonable time frame) and at different heights within the lowest 1000 m AGL. It soon became evident that those trajectories starting along the Texas coast and adjacent interior were more likely to pass in close proximity to the Dakotas and/or the southern MRB than others starting farther to the east. The matrix was then reduced in extent to include only those starting points located within the preferred source region. Next the horizontal resolution was increased to 0.5° and parcels were again released at different heights and times until trajectories were found that reached the Dakotas and/or the southern MRB. A similar methodology was adopted to identify backward trajectories starting over the southern MRB, except there the initial extent of the matrix was limited to those contiguous areas that received more than 50 mm of precipitation.

c. Limitations of the trajectory calculations

The exact path of a trajectory is sensitive to the initial conditions, uncertainties contained in the wind fields, and interpolation errors. Rolph and Draxler (1990) investigated the sensitivity of trajectories to changes in temporal and spatial density of meteorological data during a wide range of synoptic conditions. The trajectories were calculated using analysis and forecast fields from the NGM model. For horizontal grid resolutions of 90 and 180 km, they found that the trajectories were most sensitive to changes in temporal resolution of the input data fields. Specifically, using a grid resolution of 90 km and a temporal resolution of 2 h, mean relative horizontal transport deviations of approximately 5% of the trajectory length were observed after 96 h. Using a temporal resolution of 6 h, the mean relative horizontal error increased to 15%.

Stohl (1998) concluded that position errors of close to 20% of the travel distance are typical for trajectories calculated using analysis fields in data-rich areas (e.g., North America), with errors of up to 30% or more expected for trajectories calculated using forecast data. In contrast, Reiff et al. (1986) verified trajectory calculations for an African dust plume using wind analyses produced at the European Centre for Medium-Range Weather Forecasts and found that the errors were at most 200 km (∼7%) for a 3000-km trajectory. For a comprehensive review of error sources in trajectory calculations, the reader is referred to Stohl (1998).

A methodology that has been employed to test the sensitivity of trajectory calculations to the input data is the ensemble approach. For example, Draxler (2003) devised an ensemble trajectory analysis technique that can generate multiple simulations from a single meteorological dataset. Each ensemble member is computed from the same source location, but during the calculation the meteorological grid is offset by an amount determined by the user. For example, Draxler (2003) used an offset of ±1 grid point in the horizontal direction and ±250 m in the vertical direction, resulting in an ensemble of 27 members. The resulting ensemble of trajectories is useful for illustrating the uncertainties in the trajectory calculations that can be introduced by limitations in the spatial resolution of the meteorological data. In section 4, we will present an example of the ensemble technique. For more details on the rationale and methodology used in the ensemble technique, the reader is referred to Draxler (2003).

It is important to note that the “parcels” referred to in this study do not represent the hypothetical parcels used in buoyancy theory. In other words, mixing with the surrounding air does occur (see section 4b) and some of the moisture is removed by precipitation. Consequently, the mixing ratios specified along the trajectory represent the moisture content at particular points and times, and do not suggest that all of the moisture present in the air originated from the GOM. One could employ isotopic tracer techniques (e.g., Salati et al. 1979) to determine whether water molecules entering the rain systems over the southern MRB originated over the GOM. However, the properties of the rainfall measured over the southern MRB were not measured, making an isotopic analysis impossible.

4. The 22–23 June 1993 case

a. Background and synoptic setting

Between 22 and 23 June 1993, a rainstorm produced 50 mm or more of precipitation over an area of approximately 136 000 km2 (∼30% of the southern MRB) in less than 48 h (Table 1). Figure 2a shows that the highest rainfall amounts (up to 140 mm) were recorded over the Swan Hills, located about 150 km northwest of Edmonton, Alberta (Fig. 1).

Figure 3a shows the dramatic increase in the streamflow rates following the event. Specifically, the runoff from the rainstorm increased the flow of rivers in the Peace River basin by 300%–400% in the days following the event. Flow rates at locations collocated with the heavy rainfall (P1 and P2) peaked approximately three days after the rain started, while this surge of water was visible about 500 km downstream at P3 some nine days later. The secondary peaks evident at P1 and P2 in Fig. 3a are associated with a separate rainfall event that occurred shortly after the 22–23 June event. These data suggest that the effects of this rainfall event were noticeable at locations far removed from the region that experienced heavy rainfall.

The high rainfall amounts observed between 22 and 23 June 1993 were produced by a classic synoptic-scale setting for heavy summer rainfall events over central Alberta. Specifically, a 500-hPa cutoff low (5460 gpm) was located just off the coast of British Columbia on the morning of 22 June 1993, and over the next 36 h the cutoff low slowly migrated northeastward. As the system approached the continental divide, rapid lee cyclogenesis occurred over south-central Alberta, with the central pressure of the surface low decreasing from 1006 hPa at 1200 UTC 21 June to 988 hPa at 0000 UTC 23 June. Figures 4a and 4b depict the 500-hPa and surface circulation pattern at 1200 UTC 23 June 1993. At this time, heavy rain was falling over the southern MRB, the central pressure of the surface low over Alberta was 986 hPa, and the heights at the center of the 500-hPa cutoff low were 5440 gpm.

b. Trajectory analysis

An example of a forward trajectory for the 1993 extreme rainfall event is shown in Fig. 5a. Hereafter, we refer to this type of trajectory as “quasi-continuous” to indicate that the trajectories are not strictly continuous in space and time. This is evidenced by the tight loops that are sometimes found along the trajectories (e.g., Figs. 5c,e). Inspection of the trajectory data indicates that these loops coincide with times when the winds were light and/or precipitation events were occurring.

Figure 5a shows how, over a period of about 4 days, very moist boundary layer air (initially 20 g kg−1) from the GOM was advected from the Texas coast northward through the Texas panhandle to the Midwestern United States. In a study of warm-season moisture transport over the United States, Schubert et al. (1998) referred to this region of preferred moisture transport as the “Texas corridor.” The arrival of the modified GOM moisture over Nebraska coincided with the formation of a strong southerly flow over the northern Great Plains and Saskatchewan in response to rapid lee cyclogenesis over south-central Alberta (Fig. 4b).

We refer to “modified” moisture content because Fig. 5b shows that there is a decrease in the mixing ratio with time as the parcel progresses northward. At constant pressure, the only means of reducing the mixing ratio is to either remove moisture though precipitation processes or by turbulent mixing with drier air. The data along the trajectories suggest that the diurnal fluctuation in the low-level mixing ratio might be attributed to turbulent mixing with dry elevated mixed-layer air (Lanicci and Warner 1991) that is often present over the high plains during the warm season, and/or rainfall associated with vertical ascent. Both of these processes would result in a reduction in the mixing ratio. Nevertheless, the mean precipitable water content (PW) over the southern MRB, calculated from 12-hourly soundings released from Stony Plain (see Fig. 1), for the duration of this event was 21 mm. This is 40% higher than the long-term mean PW observed at this time of the year. Moreover, the 1200 UTC sounding released from Stony Plain on 23 June 1993 showed a local θe maximum (∼316 K) between 700 and 600 mb, reflecting the influx of very moist air into this layer.

Comparison of 6-hourly surface and upper-air analyses and coincident trajectory locations indicated that over a period of about 48 h the strong cyclonic flow around the cutoff-low system drew the modified GOM moisture (>11 g kg−1) from the Dakotas northward to central Saskatchewan and finally westward over the southern MRB. The total time required to transport the modified GOM moisture to the southern MRB, a distance of approximately 3500 km, was 6 days. Figure 5a shows only one of several quasi-continuous trajectories that could be followed from the vicinity of the GOM to the southern MRB (not shown). Although these other trajectories started at different times, heights, and locations, their properties were similar in terms of location, elevation, duration, and moisture content to the trajectory shown in Fig. 5.

It is clear from the short time taken to transport the moisture from the GOM to the northern plains that the Great Plains low-level jet (LLJ) likely played a key role in the northerly transport of GOM moisture during this event. The Great Plains LLJ is frequently observed over the Midwestern United States during spring and summer (Helfand and Schubert 1995). Consequently, the Great Plains LLJ has a pronounced effect on moisture transport and precipitation over the central United States. Specifically, moisture budget analyses for the continental United States indicate that strong poleward transport of GOM moisture exists in the vicinity of the Great Plains LLJ (Schubert et al. 1998; Liu and Stewart 2003). The northerly transport of GOM moisture over the Great Plains typically peaks in June and July, and at this time also extends as far north as the Dakotas and Wisconsin (Liu and Stewart 2003). Higgins et al. (1997) showed that the frequent northward transport of GOM moisture to the Midwestern United States by the Great Plains LLJ accounts for a significant portion (up to 45%) of the regional moisture transport in that region. Likewise, Helfand and Schubert (1995) showed that the Great Plains LLJ is responsible for transporting almost one-third of all of the moisture that enters the continental United States annually.

Higgins et al. (1997) found that Great Plains LLJ events are associated with enhanced warm-season precipitation over the north-central states and Great Plains. Also, Fritsch et al. (1986) found that between 30% and 70% of the warm-season rainfall over the Midwestern United States falls from nocturnal mesoscale convective systems (MCSs). MCSs tend to develop as a result of, and are sustained by, the lift and convergence present at the northern end of an LLJ (Walters and Winkler 2001). Moreover, the most intense precipitation episode during the 1993 flood over the upper Mississippi River basin was associated with a sustained strong southerly low-level jet (Arritt et al. 1997).

A noteworthy observation from Fig. 5b is the ascent of moist low-level air over Saskatchewan and the southern MRB as it was lifted into the cutoff low during the last 36 h of its trajectory. The water vapor mixing ratio of the air prior to this rapid ascent was 11 g kg−1. The combination of sustained moderate vertical ascent and a continuous supply of moist air over the southern MRB favored significant rainfall over the basin. Indeed, continuous rainfall was observed to last up to 48 h at many locations, while the average rainfall for all of those stations in the southern MRB that reported precipitation was approximately 54 mm.

The much higher rainfall amounts (>120 mm) evident over the Swan Hills in Fig. 2a could possibly be attributed to strong orographic lift on the hills’ windward slopes. Lin et al. (2001) investigated heavy orographic rainfall events over North America, Europe, and Asia. They proposed an index, U(∂h/∂x)q, where U is the low-level flow velocity perpendicular to the mountain range, ∂h/∂x is the mountain slope parallel to the basic flow, and q is the low-level water vapor mixing ratio, to help predict the occurrence of heavy orographic rainfall. The Swan Hills rise to a maximum height of 1300 to 1400 m above mean sea level, and the northerly facing slopes have a gradient of about 0.010. Soundings released from Stony Plain showed that the 850-hPa winds, which closely correspond to the height of the Swan Hills, were northerly and varied between 20 and 25 m s−1 for approximately 24 h during this event. The observed 850-hPa mixing ratio during this time was near 6 g kg−1. The peak value of the heavy orographic rainfall index over the Swan Hills for this event was near 2.0 and translates into a rainfall rate of approximately 3.5 mm h−1, or 80 mm in 24 h. Thus, the strong orographic lift of moist air on the windward slopes of the Swan Hills would have undoubtedly acted to enhance the rainfall amounts there.

c. Ensemble trajectories

The technique of Draxler (2003) was used to calculate an ensemble of backward trajectories from the southern MRB for this event. The ensemble trajectories shown in Fig. 6 were calculated using a horizontal offset of about 90 km and a vertical offset of approximately 125 m. The trajectories diverge with time, with three preferred source regions emerging after 162 h. Nine of the 27 ensembles could be traced back to the vicinity of the GOM and 11 of the ensembles to the Pacific. Six of the trajectories were terminated when they intercepted the ground over the high plains. The mean mixing ratio along the trajectories originating over the GOM and high plains differed significantly from those originating over the Pacific. Specifically, the average mixing ratio of the GOM parcels (e.g., see parcel G in Fig. 6) and those from the High Plains prior to rapid ascent were almost 13 and 6 g kg−1, respectively. In contrast, the average mixing for those parcels originating over the Pacific (see parcel P) was approximately 5 g kg−1. Moreover, during the final 40 h of the trajectories, the parcel originating near the GOM (G) ascended at a mean rate of 3 cm s−1, versus only 1.5 cm s−1 for the parcel originating over the Pacific (P).

Thus, while some trajectories were found to originate from the Pacific, they contained only about a third of the moisture compared to those trajectories originating over the GOM and did not ascend as fast. These data also suggest the modified GOM moisture may have been responsible for a significant contribution to the moisture feeding the rainstorm.

5. The 18–19 June 1996 case

a. Background and synoptic setting

This extreme rainfall event produced more than 100 mm of rain over portions of the southern MRB in less than 48 h. Approximately 57 000 km2 (∼12% of the southern MRB) received 50 mm or more of rain (Table 1). Figure 2b shows that the highest rainfall amounts were recorded over the Swan Hills, where up to 140 mm of rain was measured. The average accumulated rainfall of those stations in the southern MRB that reported rain was approximately 47 mm. Hourly station data indicate that the longest observed period of continuous rain lasted from 40 to 45 h.

The heavy rainfall produced localized flooding and 200 people had to be evacuated from the mouth of the Driftpole River on Slave Lake. Figure 3b indicates that the streamflow of rivers in the Peace River basin increased significantly in the days following the event, with streamflow rates peaking at 300%–500% of their prestorm values. In the immediate vicinity of the heavy rainfall (P1 and P2) the flow peaked two days following the onset of rain, while farther downstream (P3) the flow peaked six to seven days after the heavy rains started.

The synoptic setting shown in Figs. 4c and 4d for this event is similar to that described for the 1993 case study. On 17 June 1996, a deep cutoff low (5480 gpm) approached the British Columbia coast. Over the next 36 h, the low tracked steadily northeastward, spawning a deep surface low in the lee of the continental divide over central Alberta on the morning of 18 June. The system reached its peak intensity near 1200 UTC 19 June (Figs. 4c,d). At this time, the central pressure of the surface low over Alberta was 994 hPa and the heights at the center of the 500-hPa cutoff low were near 5440 gpm. By 0000 UTC 20 June, the system had crossed the Alberta/Saskatchewan border and had begun to weaken.

b. Trajectory analysis

Trajectory analyses of this event revealed two primary modes of moisture transport from the GOM. Figure 5c shows an example of a quasi-continuous trajectory that was very similar to the trajectory shown in Fig. 5a for the 1993 event. In particular, very moist boundary layer air (∼12 g kg−1) over southeastern Texas was transported northward to South Dakota by the Great Plains LLJ in about seven days. The quasi-continuous nature of the trajectory is illustrated by the two tight loops evident along its length. According to the U.S. unified precipitation dataset (from the Climate Diagnostics Center; http://www.cdc.noaa.gov/), the loop over Kansas was associated with a rainfall event that produced up to 25 mm of rain on 12 June. Similarly, the loop over South Dakota was located in close proximity to a band of heavy rainfall (maximum amounts near 50 mm) observed along the border of North and South Dakota on 15 June.

During the final three days of the trajectory, modified GOM moisture over South Dakota was first advected rapidly northward over central Saskatchewan (in response to the lee cyclogenesis over Alberta) and then westward by the cyclonic flow around the northern periphery of the surface and upper-air lows. The 0000 UTC Stony Plain sounding released on 19 June 1996 showed a local θe maximum (∼313 K) near 520 mb, reflecting the influx of moist air at this level. The total time required for the trajectory that originated near the GOM to reach the southern MRB was about 10 days.

Dirmeyer and Brubaker (1999) found that the northward transport of GOM moisture over the Mississippi River basin during the summer months often occurs in a stepwise fashion. Specifically, they found that most of the GOM moisture is transported northward incrementally, with moisture often first falling over the lower Mississippi basin, then evaporating before it is advected northward and falls again as precipitation. This process repeats itself, with the average contribution of GOM moisture to precipitation events decreasing from south to north over the basin.

Our trajectory analyses revealed a similar stepwise transport of moisture, and an example of a stepwise trajectory is shown in Fig. 7. This trajectory involved three discrete stages: during the first stage, the Great Plains LLJ transported moist GOM air (∼13 g kg−1) from southern Texas to Kansas (Fig. 7a). As was mentioned previously, the loop over Kansas was associated with a rainfall event that produced up to 25 mm of rain on 12 June. The second stage was completed after about six days, when the trajectory reached the Dakotas. The arrival of the trajectory (which originated near the GOM) over the Dakotas on 15 June coincided with the formation of an MCS over southeastern North Dakota that produced up to 50 mm of rain in some areas. The importance of evapotranspiration in moistening the planetary boundary layer has been well documented (e.g., Brubaker et al. 2001; Szeto 2002). It is argued, therefore, that during the short break before the initiation of the third stage of the moisture transport on 16 June, evapotranspiration from the very moist underlying surface maintained the high moisture content in the boundary layer above southeastern North Dakota. This is supported by the well-mixed boundary layer moisture (∼11 g kg−1) that was measured by the soundings released from Bismark, North Dakota, on 16 June.

Figure 7c depicts the third stage of the moisture transport. Within 24 h of the 15 June heavy rainfall event over North Dakota, the deepening low over Alberta forced a strong southerly flow over the northern high plains and Saskatchewan. During the following 60 h or so, the strong cyclonic flow around the cutoff-low system transported moist boundary layer air from North Dakota northward over Saskatchewan, and then westward over the southern MRB. The time required for the trajectory, which originated near the GOM, to reach the southern MRB was 10 days.

As was the case for the 1993 event, rapid ascent occurred during the last 24 h of the trajectory (see Fig. 7d). The mixing ratio of the air before rapid ascent was 11.5 g kg−1. Examination of the Stony Plain soundings released during this event indicated that the high rainfall amounts over the Swan Hills could be attributed to the strong and persistent upslope flow (northerly LLJ of 20 to 25 m s−1).

6. The 28–29 July 2001 case

a. Background and synoptic setting

Figure 2c shows that portions (32 000 km2 or 7%) of the southern MRB received 50 mm or more rain during this event, with over 100 mm being recorded along the extreme southern periphery of the Athabasca River basin. The region affected by heavy rainfall during this event represents a significantly smaller area than was observed in the other two case studies presented in this paper. However, between 20 and 21 July a rainstorm produced about 60 mm of rain over portions of west-central Alberta. This was followed by another rainstorm between 23 and 25 July 2001, which produced up to 90 mm of rain over most of the area affected by the 28–29 July event. The first two rainstorms would thus have saturated the ground and this would have enhanced the runoff during the 28–29 July event. This enhanced flow is reflected in the streamflow data shown in Fig. 3c, which shows the large influx of water into the Athabasca River basin. In particular, streamflow rates at A1 and A2 increased by 20%–60% above their prestorm values in the days following the 28–29 July event. The peak streamflow in the lower reaches of the basin (A3) was observed some five days after the rain started.

It is evident from Figs. 4e and 4f that the weather system responsible for producing the heavy rainfall was slightly weaker than observed for the other two cases. During the morning of 28 July 2001, a 500-hPa cutoff low (5600 gpm) crossed the southern British Columbia coast. At the same time, a surface low developed over southern Alberta. During the next 36 h the 500-hPa cutoff low tracked northeastward crossing the Alberta/Saskatchewan border around 1800 UTC 29 July. During this period, the surface low moved slowly northward along the Alberta/Saskatchewan border and deepened from 1006 to 992 hPa. Figures 4e and 4f show the 500-hPa and surface maps valid for 1200 UTC 29 July. At this time, the minimum surface pressure was about 994 hPa and the heights at the center of the 500-hPa cutoff low were near 5560 gpm.

b. Trajectory analysis

Trajectory analyses for this event identified both quasi-continuous and stepwise trajectories that transported moisture from the GOM to the southern MRB. Figure 8 depicts how modified GOM moisture was transported northward to the southern MRB along a stepwise trajectory. The arrival of the trajectory over western Dakota during the evening of 26 July coincided with the formation of an MCS that produced over 100 mm of rain and caused widespread flooding. Approximately 24 h following this rainstorm, moist low-level air (>11 g kg−1) originating from above the rain-soaked soil of western North Dakota was advected northward over Saskatchewan and arrived over the southern MRB some 36 h later. As was observed for the other two extreme rainfall events, the deep upper-air and surface lows caused a strong cyclonic flow throughout the lower troposphere over Saskatchewan and the southern MRB.

An example of a quasi-continuous trajectory computed for this event is shown in Figs. 5e and 5f, which depict how boundary layer moisture (13 g kg−1) was transported from the Texas coast northward to the Midwestern United States over a period of about five days. During this entire time, the trajectory remained in the boundary layer. The loop in the trajectory observed over the tristate boundaries of South Dakota, Nebraska, and Iowa on 23 and 24 July 2001 coincides with a heavy rainfall event, as indicated by hourly Next Generation Weather Radar (NEXRAD) national mosaic images (available online at http://www.ncdc.noaa.gov/oa/radar/radarresources). Radar data also indicated heavy precipitation collocated with the loop over eastern Montana on 26 and 27 July 2001.

In the final 48 h of the trajectory, modified GOM moisture was drawn rapidly northward by the strong cyclonic flow that had developed ahead of the cutoff surface and upper-air lows over Alberta. Rapid vertical ascent occurred during the last 24 h of both the quasi-continuous and stepwise trajectories as the air was ingested into the system. The mixing ratio of the air before rapid ascent was 11 g kg−1, and was then depleted to about 3 g kg−1 as the parcel underwent rapid ascent and precipitation was initiated. The presence of this abnormally moist air was reflected in the 1200 UTC Stony Plain sounding released on 29 July 2001, which showed a local θe maximum (∼323 K) near 750 mb. The total time taken to transport the modified GOM moisture from the Texas coast to the southern MRB was about nine days.

c. Ensemble trajectories

The ensemble trajectories shown in Fig. 9 were calculated using the same horizontal and vertical offsets as used for the 1993 event. The trajectories diverge significantly with time along the 240-h simulation, especially the high-altitude trajectories originating over the North Pacific. Nine of the 27 ensembles could be traced back to the high latitudes (45° to 60°N) over the Pacific, 6 to the middle latitudes (25° to 45°N) over the North Pacific, 6 to the vicinity of the GOM and the high plains, and 5 to the Arctic (>60°N). The mean mixing ratio along the trajectories originating over the GOM and high plains differed significantly from those originating over the Pacific and Arctic. Specifically, the average mixing ratio along the GOM trajectories and those from the high plains prior to rapid ascent were between 11 and 13 g kg−1, respectively. In contrast, the average mixing ratios calculated along those trajectories originating over the Arctic and Pacific were approximately 3 and 5 g kg−1, respectively. Their lower mixing ratios could, in part, be attributed to the fact that for most of the time they were located above the boundary layer. The two trajectories that were traced back to northern Hudson Bay after 240 h are interesting, because despite originating at high latitudes, the mean mixing ratio along the trajectories prior to rapid ascent was approximately 9 g kg−1. When the trajectories were over Hudson Bay they had a mixing ratio of only 4 to 5 g kg−1. However, as the parcels traveled southwestward their moisture content increased significantly, and by the time they were over Nebraska their mixing ratios had increased to 13 g kg−1. The parcels were then transported northward, and during the final 30 h were lifted rapidly as they were ingested into the cutoff low over Alberta.

Therefore, although some trajectories were found to originate from the Pacific and the Arctic, they contained less than half of the moisture than those trajectories originating over the GOM and high plains. The ensemble trajectories for the 2001 rainstorm suggest that the modified GOM moisture was probably a significant source of moisture.

7. Discussion and conclusions

The large-scale transport of low-level moisture has been investigated for three meso-α-scale extreme rainfall events over the southern MRB. It is evident from Fig. 4 and statistics presented in Table 1 that there was a striking similarity between the synoptic-scale circulation patterns (both at the surface and at 500 hPa) that were observed over North America during these events. Specifically, all three events were associated with a 500-hPa cutoff low and a deep surface low over central Alberta. While June and July are favored months for cutoff lows over the southern MRB, rarely do these systems produce over 100 mm of rain as they traverse the basin. We are, therefore, dealing with a special subset of rainstorms that make a significant contribution (25%) to the annual moisture budget for a small portion (10%–30%) of the southern MRB in less than 48 h. For two of the three events, the heaviest rain was observed over the Swan Hills. One possible reason for this regional maximum was a persistent and moist northerly LLJ (>20 m s−1) that produced strong upslope flow over the north-facing slopes of the Swan Hills. Our research identified the GOM as a possible source of low-level moisture for all three extreme rainfall events. This corroborates the hypothesis of Liu and Stewart (2003) that the GOM is a source of moisture over the Saskatchewan River basin and the southern MRB during the early summer.

The similarities between the synoptic-scale circulation patterns were reflected in the trajectory analyses computed for each case study (Fig. 5). Specifically, the modified GOM moisture followed very similar trajectories and required about the same amount of time (6 to 10 days) to reach the southern MRB. This represents a 3500-km poleward transport of moist, subtropical air in 10 days or less. It is remarkable that moist air originating over subtropical waters can feed rainstorms in the high latitudes and ultimately flow into the Arctic Ocean.

Our trajectory analyses identified two types of trajectories along which GOM moisture was transported to the southern MRB. We refer to the first type as a quasi-continuous trajectory and the second as a stepwise trajectory. A schematic of the proposed conceptual model for the two types of moisture transport for extreme rainfall events over the southern MRB is presented in Fig. 10. A critical component of both types of trajectories was the Great Plains LLJ. Research has shown that the Great Plains LLJ plays a key role in the summer precipitation and hydrology of the central United States. In particular, the Great Plains LLJ is an essential mechanism for transporting GOM moisture over the central United States during the spring and summer months, and this transport in turn has a marked effect on precipitation patterns over the Midwestern United States (Helfand and Schubert 1995; Higgins et al. 1997; Schubert et al. 1998; Liu and Stewart 2003).

In the case of quasi-continuous trajectories, the GOM moisture was initially transported northward to the northern plains of the United States by the Great Plains LLJ through the so-called Texas corridor (Schubert et al. 1998). The high frequency of LLJ events over the Great Plains of the United States accounts for a significant portion of the regional moisture transport in this region during the warm season. What is very uncommon, however, is for the arrival of modified GOM moisture over the central and northern Great Plains to coincide with lee cyclogenesis over Alberta that is associated with a deep 500-hPa cutoff low. Under these circumstances, the low-level moisture that was advected over the northern plains by the Great Plains LLJ was caught up in the strong southerly flow developing ahead of the cutoff low located over central Alberta. This moisture was then transported northward over Saskatchewan and finally westward over the southern MRB by the cyclonic flow within the cutoff-low system. The airflow during the last 36 h or so in our three case studies appears to be very similar to that documented by Bierly and Winkler (2001) in their study of early developing Colorado cyclones. Specifically, our trajectories appear to bear close resemblance to the trajectory of their cyclonically turning moist airstream (CMA). According to Bierly and Winkler, the CMA is characterized by “its distinct cyclonic curvature and likely trajectory over a moisture source.” The arrival of very moist air in the midlevels of the troposphere is supported by local θe maxima present in Stony Plain soundings between 750 and 500 mb for all three rainfall events. For the cases presented herein, it typically took those quasi-continuous trajectories originating in the vicinity of the GOM a week to 10 days to reach the southern MRB.

To illustrate the importance of the Great Plains LLJ in transporting the modified GOM moisture, we computed the 850-hPa meridional wind anomalies (from NCEP reanalysis data; Kalnay et al. 1996) from the time the trajectories started until just before the onset of rapid ascent. The winds at 850 hPa are deemed to correspond closely to the elevation where the LLJ reaches its peak intensity over the high plains (Pan et al. 2004). Figure 11 indicates that an enhanced southerly flow at 850 hPa was evident over the Midwestern United States prior to the 1993 and 2001 rainfall events over the southern MRB. The anomaly maps also suggest that the Great Plains LLJ and associated moisture transport extended farther north than is typically observed at this time of the year. The anomalously strong southerly low-level flow (>5 m s−1) over Saskatchewan was related to the formation of a deep surface low (<1000 hPa) over Alberta. It follows from Fig. 11 that the enhanced southerly flow at 850 hPa over the Great Plains and Saskatchewan was probably critical in facilitating the transport of modified low-level, subtropical moisture from the GOM to the southern MRB.

While the requirements for stepwise trajectories are less restrictive than for quasi-continuous trajectories, they still require a special set of atmospheric conditions. For the case studies discussed herein, the trajectories originating in the vicinity of the GOM arrive over the Dakotas within a day or two before rapid lee cyclogenesis over central Alberta. Moreover, the arrival of the trajectories appears to coincide with the formation of an MCS that produces widespread and heavy rain over the Dakotas. While the climatological maximum of the Great Plains LLJ is located south of the Dakotas over Kansas (Bonner 1968), Mitchell et al. (1995) noted that the maximum frequency of the stronger low-level jets and associated precipitation extended farther north (over Nebraska) than for weaker low-level jets. Indeed, 850-hPa meridional wind anomaly maps computed immediately prior to the heavy rainfall events over the Dakotas on 15 June 1996 and 26 July 2001 show that the low-level southerly winds over the Dakotas were stronger than their long-term averages. The final stage of the moisture transport was realized when the strong cyclonic low-level flow around the deepening cutoff-low system over Alberta advected the modified GOM moisture over the southern MRB.

In conclusion, the trajectory analyses for the three cases presented herein suggest that rapid lee cyclogenesis over Alberta (associated with a 500-hPa cutoff low) and the Great Plains LLJ acted in unison to transport modified subtropical moisture from the GOM to the southern MRB. Given the importance of meso-α-scale extreme rainfall events on the hydrological cycle of the MRB, future work should focus on conducting a detailed study of the dynamics responsible for producing the heavy rainfall, as well as developing analog techniques to assist forecasters in identifying cases when the antecedent large-scale flow has the potential to produce extreme rainfall over the southern MRB. Also of interest would be to investigate the moisture sources on occasions when the synoptic pattern was similar to the events presented in this paper, but the rainfall amounts were less than 100 mm.

Acknowledgments

This work was supported by a Natural Sciences and Engineering Research Council (NSERC) Collaborative Special Grant in support of MAGS. The authors acknowledge the NOAA/Air Resources Laboratory (ARL) for providing the HYSPLIT transport and dispersion model, and the U.S. unified precipitation data was provided by the NOAA–CIRES Climate Diagnostics Center from their Web site at http://www.cdc.noaa.gov/. The authors thank Alberta Environment for providing rainfall data and streamflow data. Thanks also to Nicole Hopkins for creating the maps.

REFERENCES

  • Arritt, R. W., , Rink T. D. , , Segal M. , , Todey D. P. , , Clark C. A. , , Mitchell M. J. , , and Labas K. M. , 1997: The Great Plains low-level jet during the warm season of 1993. Mon. Wea. Rev., 125 , 21762192.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, G. D., , and Bosart L. F. , 1989: A 15-year climatology of Northern Hemisphere 500 mb closed cyclone and anticyclone centers. Mon. Wea. Rev., 117 , 21422164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bierly, G. D., , and Winkler J. A. , 2001: A composite analysis of airstreams within cold-season Colorado cyclones. Wea. Forecasting, 16 , 5780.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bjornsson, H., , Mysak L. A. , , and Brown R. D. , 1995: On the interannual variability of precipitation and runoff in the Mackenzie drainage basin. Climate Dyn., 12 , 6776.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bonner, W. D., 1968: Climatology of the low-level jet. Mon. Wea. Rev., 96 , 833850.

  • Brubaker, K. L., , Dirmeyer P. A. , , Sudradjat A. , , Levy B. S. , , and Bernal F. , 2001: A 36-yr climatological description of the evaporative sources of warm-season precipitation in the Mississippi River basin. J. Hydrometeor., 2 , 537557.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chung, Y., , Hage K. D. , , and Reinelt E. R. , 1976: On lee cyclogenesis and airflow in the Canadian Rocky Mountains and the East Asian Mountains. Mon. Wea. Rev., 104 , 879891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., , and Brubaker K. L. , 1999: Contrasting evaporative moisture sources during the drought of 1988 and the flood of 1993. J. Geophys. Res., 104 , 1938319397.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., , 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
  • Draxler, R. R., 2003: Evaluation of an ensemble dispersion calculation. J. Appl. Meteor., 42 , 308317.

  • Draxler, R. R., , and Hess G. D. , 1997: Description of the HYSPLIT_4 modeling system. NOAA Tech. Memo. ERL ARL-224, 24 pp.

  • Draxler, R. R., , and Hess G. D. , 1998: An overview of the HYSPLIT_4 modelling system for trajectories, dispersion, and deposition. Aust. Meteor. Mag., 47 , 295308.

    • Search Google Scholar
    • Export Citation
  • Fink, A. H., , and Knippertz P. , 2003: An extreme precipitation event in southern Morocco in spring 2002 and some hydrological implications. Weather, 58 , 377387.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fritsch, J. M., , Kane R. J. , , and Chelius C. R. , 1986: The contribution of mesoscale convective weather systems to the warm-season precipitation in the United States. J. Appl. Meteor., 25 , 13331345.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Helfand, H. M., , and Schubert S. D. , 1995: Climatology of the simulated Great Plains low-level jet and its contribution to the continental moisture budget of the United States. J. Climate, 8 , 784806.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , Yao Y. , , Yarosh E. S. , , Janowiak J. E. , , and Mo K. C. , 1997: Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States. J. Climate, 10 , 481507.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hoke, J. E., , Phillips N. A. , , Dimego G. J. , , Tuccillo J. J. , , and Sela J. G. , 1989: The regional analysis and forecast system of the national meteorological center. Wea. Forecasting, 4 , 323334.

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

  • Kanamitsu, M., 1989: Description of the NMC global data assimilation and forecast system. Wea. Forecasting, 4 , 335342.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Keim, B. D., 1998: Record precipitation totals from the coastal New England rainstorm of 20–21 October 1996. Bull. Amer. Meteor. Soc., 79 , 10611068.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lackmann, G. M., , Gyakum J. R. , , and Benoit R. , 1998: Moisture transport diagnosis of a wintertime precipitation event in the Mackenzie River basin. Mon. Wea. Rev., 126 , 668692.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lanicci, J. M., , and Warner T. T. , 1991: A synoptic climatology of the elevated mixed-layer inversion over the southern Great Plains in spring. Part I: Structure, dynamics, and seasonal evolution. Wea. Forecasting, 6 , 181197.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y., , Chiao S. , , Wang T. , , Kaplan M. , , Michael L. , , and Weglarz R. P. , 2001: Some common ingredients for heavy orographic rainfall. Wea. Forecasting, 16 , 633660.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, J., , and Stewart R. E. , 2003: Water vapor fluxes over the Saskatchewan River basin. J. Hydrometeor., 4 , 944959.

  • Louie, P. Y. T., , Hogg W. D. , , MacKay M. D. , , Zhang X. , , and Hopkinson R. F. , 2002: The water balance climatology of the Mackenzie River basin with reference to the 1994/1995 water year. Atmos.–Ocean, 49 , 159180.

    • Search Google Scholar
    • Export Citation
  • Mitchell, M. J., , Arritt R. W. , , and Labas K. , 1995: A climatology of the warm season Great Plains low-level jet using wind profiler observations. Wea. Forecasting, 10 , 576591.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Novak, M. J., , Bosart L. F. , , Keyser D. , , Wasula T. A. , , and LaPenta K. D. , 2002: Climatology of warm season 500 hPa cutoff cyclones and a case study diagnosis of 14–17 July 2000. Preprints, 19th Conf. on Weather Analysis and Forecasting, San Antonio, TX, Amer. Meteor. Soc., 68–71.

  • Pan, Z., , Segal M. , , and Arritt R. W. , 2004: Role of topography in forcing low-level jets in the central United States during the 1993 flood-altered terrain simulations. Mon. Wea. Rev., 132 , 396403.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulos, G. S., , Wesley D. A. , , Snook J. S. , , and Meyers M. P. , 2002: A Rocky Mountain storm. Part I: The blizzard—Kinematic evolution and the potential for high-resolution numerical forecasting of snowfall. Wea. Forecasting, 17 , 955970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Proctor, B. A., , Wang M. , , Strong G. S. , , and Smith C. D. , 1999: Atmospheric moisture budgets for MAGS. Proc. Fifth Scientific Workshop for the Mackenzie GEWEX Study (MAGS), Edmonton, AB, Canada, Environment Canada and the Natural Sciences and Engineering Research Council of Canada, 117–124.

  • Randel, D. L., , Greenwald T. J. , , Vonder Haar T. H. , , Stephens G. L. , , Ringerud M. A. , , and Combs C. L. , 1996: A new global water vapor dataset. Bull. Amer. Meteor. Soc., 77 , 12331246.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reiff, J., , Forbes G. S. , , Spieksma F. T. M. , , and Reynders J. J. , 1986: African dust reaching northwestern Europe: A case study to verify trajectory calculations. J. Climate Appl. Meteor., 25 , 15431567.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reuter, G. W., , and Nguyen C. D. , 1993: Organization of cloud and precipitation in an Alberta storm. Atmos. Res., 30 , 127141.

  • Rolph, G. D., , and Draxler R. R. , 1990: Sensitivity of three-dimensional trajectories to the spatial and temporal densities of the wind field. J. Appl. Meteor., 29 , 10431054.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Salati, E., , Dall’Olio A. , , Matsui E. , , and Gat J. R. , 1979: Recycling of water in the Amazon basin: An isotopic study. Water Resour. Res., 15 , 12501258.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schubert, S. D., , Helfand H. M. , , Wu C. , , and Min W. , 1998: Subseasonal variations in warm-season moisture transport and precipitation over the central and eastern United States. J. Climate, 11 , 25302555.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smirnov, V. V., , and Moore G. W. K. , 2001: Short-term and seasonal variability of the atmospheric water vapor transport through the Mackenzie River basin. J. Hydrometeor., 2 , 441452.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stewart, R. E., and Coauthors, 1998: The Mackenzie GEWEX Study: The water and energy cycles of a major North American river basin. Bull. Amer. Meteor. Soc., 79 , 26652684.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stohl, A., 1998: Computation, accuracy and applications of trajectories—A review and bibliography. Atmos. Environ., 32 , 947966.

  • Strong, G. S., , Proctor B. , , Wang M. , , Soulis E. D. , , Smith C. D. , , Seglenieks F. , , and Snelgrove K. , 2002: Closing the Mackenzie drainage basin water budget, water years 1994/95 to 1996/97. Atmos.–Ocean, 40 , 113124.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Szeto, K. K., 2002: Moisture recycling over the Mackenzie Basin. Atmos.–Ocean, 40 , 181197.

  • Ulbrich, U., , Brucher T. , , Fink A. H. , , Leckebusch G. C. , , Kruger A. , , and Pinto J. G. , 2003: The central European floods of August 2002. Part II: Synoptic causes and considerations with respect to climate change. Weather, 58 , 434442.

    • Search Google Scholar
    • Export Citation
  • Verschuren, J. P., , and Wojtiw L. , 1980: Estimate of the maximum probable precipitation for Alberta river basins. Alberta Environment, Hydrology Branch, RMD-80/1, Edmonton, Alberta, 307 pp.

  • Walters, C. K., , and Winkler J. A. , 2001: Airflow configurations of warm season southerly low-level wind maxima in the Great Plains. Part I: Spatial and temporal characteristics and relationship to convection. Wea. Forecasting, 16 , 513530.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Whittaker, L. M., , and Horn L. H. , 1981: Geographical and seasonal distribution of North American cyclogenesis, 1958–1977. Mon. Wea. Rev., 109 , 23122322.

    • Crossref
    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

A map of the Mackenzie River basin (MRB) and Saskatchewan River basin. Subbasins over the southern MRB are delineated by gray lines. SWH represents the Swan Hills, and SP represents the upper-air sounding site at Stony Plain. Points labeled A1 to A3 and P1 to P3 represent the locations of streamflow gauges discussed in the text.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 2.
Fig. 2.

Forty-eight-hour accumulated rainfall (mm) for three extreme rainfall events over Alberta: (a) 22–23 Jun 1993, (b) 18–19 Jun 1996, and (c) 28–29 Jul 2001. The two polygons enclosed by the solid dark lines over northern Alberta approximate the boundaries of the Peace River and Athabasca River basins (see Fig. 1 for details).

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 3.
Fig. 3.

Streamflow measurements (m3 s−1) at selected locations in the Peace River and Athabasca River basins: (a) 22–23 Jun 1993, (b) 18–19 Jun 1996, and (c) 28–29 Jul 2001. Refer to Fig. 1 for the locations of streamflow measurement sites A1–A3 and P1–P3. Vertical tick marks indicate time of maximum streamflow associated with each event.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 4.
Fig. 4.

(left) The 500-hPa and (right) mean sea level pressure maps for the 22–23 Jun 1993, 18–19 Jun 1996, and 28–29 Jul 2001 extreme rainfall events. Contour interval for the 500-hPa heights is 60 gpm, and for the mean sea level pressure is 4 hPa. (Images provided by the NOAA–CIRES Climate Diagnostics Center Web site at www.cdc.noaa.gov)

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 5.
Fig. 5.

(left) Continuous trajectories from the Gulf of Mexico to the southern MRB for the three extreme rainfall events. Solid triangles are shown every 24 h. (right) The corresponding pressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 6.
Fig. 6.

(top) Backward ensemble trajectories calculated from the southern MRB for the 22–23 Jun 1993 extreme rainfall event. Trajectories labeled P and G originate from the North Pacific Ocean and the Gulf of Mexico, respectively. (bottom) The evolution in height of the trajectories (hPa).

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 7.
Fig. 7.

(left) Stepwise trajectory for the Jun 1996 extreme rainfall event. Solid triangles are shown every 12 h. (right) The correspondingpressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 8.
Fig. 8.

(left) Stepwise trajectory for the Jul 2001 extreme rainfall event. Solid triangles are shown every 12 h. (right) The correspondingpressure (light line) and water vapor mixing ratio (dark line) along each trajectory; dates are shown along the abscissas.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 9.
Fig. 9.

Same as for Fig. 6, except for backward ensemble trajectories calculated from the southern MRB for the 28–29 Jul 2001 extreme rainfall event.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 10.
Fig. 10.

Schematic of the proposed conceptual model that describes the transport of low-level moisture from the Gulf of Mexico to northern Alberta. Here CL refers to the 500-hPa cutoff low, SL to the surface low, and GPLLJ to the Great Plains low-level jet. The shaded area denotes the region where mesoscale convective complexes occur. The star indicates the location of heavy rainfall. The dashed line and block arrows represent an idealized trajectory followed by the modified low-level Gulf moisture.

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Fig. 11.
Fig. 11.

The 850-hPa merdional wind anomalies (m s−1) calculated for (a) 18–22 Jun 1993, (b) 10–18 Jun 1996, and (c) 20–28 Jul 2001. The anomalies were calculated with respect to the 1948–96 means. Solid and dashed lines represent positive and negative anomalies, respectively. Anomalies larger than +1 m s−1 are shaded. (Images provided by the NOAA–CIRES Climate Diagnostics Center Web site at www.cdc.noaa.gov)

Citation: Journal of Hydrometeorology 6, 4; 10.1175/JHM430.1

Table 1.

Summary of meteorological and hydrological parameters observed during three extreme rainfall events over the southern Mackenzie River basin (SMRB).

Table 1.
Save