A Process-Oriented Observational Study of Snowfall Potential in the Central United States
Chad M Gravelle
Saint Louis University
Charles E Graves
Saint Louis University
State University of New York College at Brockport
Scott M Rochette
Annual Missouri Academy of Science Meeting Missouri Western State University, 21 April 2007
Outline
• Introduction
• Description of Dataset
• Methodology/Case Selection
• Summary of Results
• Conclusions
Introduction
• Even with advances in numerical and ensemble model guidance, forecasting the axis and spatial extent of heavy snowfall still contains a large amount of uncertainty. • Gouree and Younkin 1966 • Browne and Younkin 1970 • Probabilistic forecasts are used by SPC and HPC to show the uncertainty with an event.
Dataset
• Daily 24-h snowfall amounts were obtained from the National Climate Data Center (NCDC) Cooperative Summary of the Day (COOP) collection for the period of November 1998 and March 2003. • North American Regional Reanalysis (NARR)
– 3-h dataset with 32-km resolution on 29 pressure levels
• NCDC publications of the Daily Weather Maps.
• COOP and NARR datasets were displayed using the General Meteorological Package (GEMPAK) software.
Methodology
Methodology
15 Jan 2003 16 Jan 2003
17 Jan 2003
72-h Event
Methodology
Methodology
• COOP event data for each station was objectively analyzed using a Barnes objective analysis.
– 271 x 171 grid ~ 10 km grid spacing
• Parameters for the objective analysis were chosen such that stations within 35 km had the most influence on each grid point. • Finally, a 9-pt smoother was applied to all fields before generating graphics.
Methodology
Case Categories
• • • •
Total events, n = 82 Banded (upper left), n = 42 Widespread (upper right), n = 15 Other/Multi-banded/Undefinable (lower left), n = 25
Banded Categories
• • •
Total events, n = 42 NW/SE (upper left), n = 11 W/E within 30° of major axis (upper right), n = 17 SW/NE (lower left), n = 14
•
Length/Width Example
Banded Length Statistics
NW/SE AVG W/E AVG SW/NE AVG NW/SE MED W/E MED SW/NE MED NW/SE SD W/E SD SW/NE SD NW/SE NORM SD W/E NORM SD SW/NE NORM SD Days 2.3 2.4 2.4 Length of > 2" Band 1389.5 1308.8 1358.4 1597.0 1353.0 1323.5 460.7 371.8 312.3 138.9 90.2 83.5 Length of > 4" Band 987.4 1139.4 1121.4 885.5 1072.0 1232.0 468.7 471.0 406.7 165.7 114.2 112.8 Length of > 6" Band 834.4 892.2 1179.5 985.0 940.5 1199.5 376.2 304.9 271.0 168.2 81.5 85.7 Length of > 12" Band none 471.3 820.3 none 455.0 780.0 none 35.4 373.6 none 20.5 152.5
Banded Width Statistics
NW/SE AVG W/E AVG SW/NE AVG NW/SE MED W/E MED SW/NE MED NW/SE SD W/E SD SW/NE SD NW/SE NORM SD W/E NORM SD SW/NE NORM SD Width of > 2" Band 220.96 256.43 283.83 206.25 271.70 232.94 122.82 81.82 148.83 37.03 19.84 39.78 Width of > 4" Band 149.79 158.56 216.50 121.37 150.57 217.43 84.94 70.81 120.06 30.03 17.17 33.30 Width of > 6" Band 93.55 119.82 165.86 79.43 122.70 170.94 44.61 55.11 91.24 19.95 14.73 28.85 Width of > 12" Band none 102.09 82.07 none 114.20 83.04 none 41.55 25.19 none 23.99 10.29
A Look Ahead...
X
A Look Ahead...
X
X
A Look Ahead...
X
X
X
A Look Ahead...
X X X X
Conclusions
• Lengths at lower snowfall thresholds often exceed the domain. • SW/NE bands are the widest, followed by E/W and NW/SE.
• Early indications suggest that the differences in width of the three orientations are not the results of sampling, but the physical processes associated with snowfall generation.
Future Research
• Using percentiles to provide forecasters with probabilistic guidance on the width of snowfall bands. • Examine the orientation and strength of processes to snowfall potential. • Expanding the period of record to include more years.
• Use the identified events as analogs of future events.
Find this, and other presentations online at:
www.eas.slu.edu/CIPS/
Questions or Comments? gravelle@eas.slu.edu