Evaluating Potential Impact of Significant East Coast Winter Storms by Analysis of Upper and

Low-Level Wind Anomalies

Neil A. StuartNOAA/NWS Albany, NY

Northeast Regional Operational Workshop VIIIAlbany, NY

1 November 2006

Considerations

NWP guidance and forecasters can still miss important atmospheric features

Recent storms illustrate NWP and forecaster busts: 25 January 2000 (Bosart 2003) 4-6 March 2001 (Grumm 2001; D’Aleo 2001) 6-7 January 2002 (Grumm 2002) 26-27 February 2004

Small changes in forecasted and observed snowfall = big difference in how people prepare for the storm, and how they are impacted

Considerations (Cont.)

Conventional NWP text and graphical guidance does not provide information on departures from normal

Climatological anomalies are calculated as standard deviations from normal Based on a 30- year climatology from 1961 through 1990

(includes the snowy decade of the 1960s) Calculated on a 2.5 by 2.5 grid (Hart and Grumm 2001)

U Wind anomalies were calculated at 850 hPa and 300 hPa, These anomalies can assist in identifying slow-moving

storms with extended periods of enhanced precipitation Real-time model and ensemble forecasts of 250 Mb

anomalies available - assumed nearly identical to 300 Mb

Why 850 Mb and 300/250 Mb?

850 Mb winds represent low level wind flow above the friction layer

Easterly winds at 850 hPa – moisture advection off the Atlantic Ocean low-level convergence enhancement of low-level frontogenesis increased precipitation production

300 to 250 Mb winds represent upper-level wind flow affecting the movement of the upper-level vorticity center

Below normal upper-level winds can signal a storm cut off from the steering flow, resulting in slower movement

Definitions Historical East Coast Winter Storm (based on

analysis of snowfall amounts and wind anomalies in 85 winter storms from 1948-2002) >12” of snow in the Carolinas and points south >18” of snow in Virginia and points north

Anomaly – Departure from normal based on a 30 year climatology (1961-1990), in units of Standard Deviations from Normal (SD)

U and V wind anomalies – Departures from normal for Westerly and Southerly winds

-U and –V wind anomalies = Departures below normal for easterly and northerly winds

4 historical storms provide overview

22-24 December 1989 Southeast U.S. case – Affected Carolinas 12” to 18” of snow in areas that normally receive <6” of snow the

entire season 5-7 March 1962 “Ash Wednesday Storm”

mid-Atlantic U.S. case – Heavy Snowfall in Virginia, Maryland, and points north

12” to 36” of snow (greatest amounts near the Appalachian Mountains) 6-8 February 1978 “Blizzard of ‘78”

Northeast U.S. case – Heavy Snowfall in southern New York and New England

2’ to 4’ of snow Over 30” of snow in Boston, Route 128 closed 3’ to 4’ of snow near Massachusetts/Rhode Island border

24-25 January 2000 “Surprise Storm” Entire East Coast case – Heavy Snowfall from Carolinas through mid-

Atlantic and northeast Widespread 12” to 24” of snow 20” of snow in Raleigh, NC

Peak 850 Mb U Wind Anomalies for Dec. 89, Mar. 62, Feb. 78 and Jan. 00

storms

Peak 300 Mb U wind Anomalies for Dec. 89, Mar. 62, Feb. 78 and Jan. 00

storms

What can we conclude?(Based on these 4 storms, and others not

shown)

850 Mb U wind anomalies <-4 SD for potential extreme snowfall

250 Mb U wind anomalies <-2.5 SD for unusually slow-moving storm

Heaviest snowfall generally within the nose of the 850 Mb –3 SD contour

Rain/Snow conversion ratios not considered

Important note

Past storm anomalies on a 2.5° X 2.5° grid, which is a resolution that is more coarse than NWP model output

Forecasted anomalies from individual NWP models can be higher than reanalysis output, due to the finer spatial resolution in the NWP model output

Snowstorms during the 2002-2003 winter illustrated differences in reanalysis and NWP model anomalies

Recent storms of 2002-2006 will illustrate differences in anomalies due to resolution, and differences in model/ensemble-depicted evolution

Christmas 2002 Snowstorm

Interior Pennsylvania through upstate New York and northern New England

Widespread 18” to 36” of snow 24” to 36+” of snow in a large

area of upstate New York

Reanalysis - Peak 850 Mb U Wind Anomaly –3.70 SDEta Model - Peak 850 Mb U Wind Anomaly <-5.5 SD

Reanalysis - Peak 300 Mb U Wind Anomaly –3.23 SDEta Model - Peak 250 Mb U Wind Anomaly <-3 SD

Another recent example – 5-7 December 2003

12” to 36” of snow northern mid-Atlantic through New England (locally @48” in northern New England)

ETA, GFS and SREF depict different anomaly values – initial and forecasted values Due to spatial resolution differences Due to differences in model-depicted evolution

ETA, GFS and SREF anomaly values all validate the anomaly thresholds for historical snowstorms

Initial and forecasted 850 hPa wind anomalies

Initial and forecasted 250 hPa wind anomalies

SREF wind anomalies from 09Z 12/6 10 member ensemble: 5 RSM, 5 ETA

15-16 December 2003

@18” of snow Lake Champlain area to northern Maine Isolated amounts to near 30” some

orographically enhanced 850 Mb anomalies exceeded the threshold for >18” of snow

Eta (-6.35 SD), GFS (-4.83), SREF (@-5 SD) 250 Mb anomalies did not reach the

threshold for a long duration storm until it reached Maine Eta (-2.49 SD), GFS (-1.90 SD), SREF (@-2.2 SD)

Carolinas Example 26-27 February 2004

12-18” of snow in central and western North and South Carolina

Very little impact in eastern North Carolina

No SREFs available, but Eta and GFS both exceeded the lower and about equaled the upper anomaly thresholds

One of the best lead times in the study

Forecasted wind anomalies

Storms and peak anomaliesEvent Maximum Observed

Snow (cm)Region Affected Peak 850 Mb Anomaly Peak upper level anomaly

Ash Wed. Storm 5-7 Mar 1962

60-90 Interior mid-Atlantic -5.49 -4.49

Blizzard of '786-7 Feb 1978

60-120 New England -5.35 -3.85

23-24 Dec 1989 30-60 Carolinas -4.11 -2.48

Blizzard of '966-8 Jan 1996

60-90 Mid-Atlantic and New England

-4.58 -2.96

Surprise Storm24-25 Jan 2000

45-60 Carolinas to New England

-4.21 -3.22

25-26 Dec 2002 60-90 Interior Northeast -3.70 (-5.9) -3.23 (-3.3)

3-4 Jan 2003 60-90 Interior Northeast -4.52 (-5.8) -1.99 (-3.3)

President's Day16-18 Feb 2003

45-75 Mid-Atlantic and Northeast

-3.57 (-5.5) -2.03 (-2.5)

6-7 Dec 2003 60-90 New England (-5.5) (-2.7)

16-17 Dec 2003 45-60 Northern New England

(-6.4) (-2.5)

Great Canada Maritime Blizzard

18-19 Feb 2004

90-120 Southeast Canada (-5.5) (-3.1)

26-27 Feb 2004 30-45 Carolinas (-5.3) (-2.7)

Blizzard of ‘0523-24 Jan 2005

45-75 New England and Southeast

Canada

-6.3 (-5.3) -2.5 (-2.5)

10-11 Feb 2005 45-60 Northern New England

-6.3 (-4.5) -3.5 (-2.5)

Two important exceptions

April Fool’s Day 1997 Storm in New England

February 2006 Storm in southern NY and NJ

Each storm was 0.5 to 1.0 SD short of the thresholds for historical storm

Exceptional frontogenesis and varying snow ratios may have contributed to extreme snowfall

850 Mb anomalies from 12 February 2006

Upper-level anomalies and frontogenesis for 12 February 2006

Final thoughts Upper-level and 850 Mb wind anomalies <-2.5 SD and <4 SD No false alarms, but POD may be impacted due to small

scale processes - frontogenesis and liquid/snow ratios MREF and SREF anomalies based on different blends of

operational models How will WRF resolve anomalies? This is just one tool for evaluating the potential for a

significant storm Model-derived forecast anomalies for 850 Mb and 250 Mb

are available in real-time at: Http://eyewall.met.psu.edu/ensembles

Thanks to Rich Grumm of NWSFO State College, PA, for his assistance in this study, and for maintaining the eyewall website with Dr. Bob Hart of Penn State University.

Thanks also to Wayne Albright, John Billet, and everyone at NWSFO Wakefield, VA for their support of this study.