FORECASTING EASTERN US WINTER STORMS Are We Getting Better and Why? Jeff S. Waldstreicher NOAA/NWS Eastern Region Scientific Services Division – Bohemia,

FORECASTING EASTERN FORECASTING EASTERN US WINTER STORMSUS WINTER STORMS

Are We Getting Better Are We Getting Better and Why?and Why?

Jeff S. WaldstreicherJeff S. WaldstreicherNOAA/NWS Eastern RegionNOAA/NWS Eastern Region

Scientific Services Division – Bohemia, NYScientific Services Division – Bohemia, NY

Northeast Regional Operational Workshop Northeast Regional Operational Workshop (NROW)(NROW)

November 4, 2004November 4, 2004

OUTLINEOUTLINE

Winter Storm Warning ProgramWinter Storm Warning Program Verification ReviewVerification Review Key Activities and DevelopmentsKey Activities and Developments Relationship Between Event Totals Relationship Between Event Totals

and Warning Performanceand Warning Performance Summary Summary

MOTIVATIONMOTIVATION Winter of 2003-2004 PerformanceWinter of 2003-2004 Performance

EASTERN REGION WIDEEASTERN REGION WIDE POD - .921 POD - .921 FAR - .321FAR - .321 Lead Time – 18.5 hoursLead Time – 18.5 hours

NORTHEASTNORTHEAST POD - .920POD - .920 FAR - .321 FAR - .321 Lead Time – 20.3 hoursLead Time – 20.3 hours

Is this indicative of a positive performance Is this indicative of a positive performance trend?trend?

If yes, what are the contributing factors?If yes, what are the contributing factors?

WINTER STORM WARNINGWINTER STORM WARNINGVERIFICATION RESULTSVERIFICATION RESULTS

1993-94 to 2003-041993-94 to 2003-04

EASTERN REGION WINTER STORM WARNINGSPOD

0.5

0.6

0.7

0.8

0.9

1

SEASON

PO

D

POD

Linear(POD)

EASTERN REGION WINTER STORM WARNINGSFAR and EVENT TOTALS

0

0.1

0.2

0.3

0.4

0.5

0.6

SEASON

FA

R

0

500

1000

1500

2000

2500

3000

# E

vent

s

FAR

#Events

Linear (FAR)

EASTERN REGION WINTER STORM WARNINGSLEAD TIME and EVENT TOTALS

8

10

12

14

16

18

20

SEASON

LT

(h

r)

0

500

1000

1500

2000

2500

3000

# E

ven

ts

LT

#Events

Linear (LT)

NORTHEASTNORTHEASTPODPOD

NORTHEAST Winter Storm WarningsAnnual Probability of Detection (POD)

0.5

0.6

0.7

0.8

0.9

1

Season

POD

0

500

1000

1500

2000

# of

Eve

nts

NORTHEAST Winter Storm WarningsPOD - 3 Year Running Mean

0.5

0.6

0.7

0.8

0.9

1

93/94

-95/96

94/95

-96/97

95/96

-97/98

96/97

-98/99

97/98

-99/00

98/99

-00-01

99/00

-01/02

00/01

-02/03

01/02

-03/04

3-Seasons

PO

D

05001000150020002500300035004000

NORTHEAST Winter Storm WarningsPOD - 2 Year Running Mean

0.5

0.6

0.7

0.8

0.9

1

2-Seasons

PO

D

050010001500200025003000

NORTHEASTNORTHEASTFARFAR

NORTHEAST Winter Storm WarningsAnnual False Alarm Ratio (FAR)

00.10.20.30.40.50.60.70.80.9

1

Season

FA

R

0

500

1000

1500

2000

# E

ven

ts

NORTHEAST Winter Storm WarningsFAR - 2 Year Running Mean

00.10.20.30.40.50.60.70.80.9

1

2-Seasons

FA

R

050010001500

200025003000

NORTHEAST Winter Storm WarningsFAR - 3 Year Running Mean

00.10.20.30.40.50.60.70.80.9

1

3-Seasons

FA

R

05001000150020002500300035004000

NORTHEASNORTHEASTT

LEAD TIMELEAD TIMENORTHEAST Winter Storm Warnings

Annual Lead Time

810121416182022

Season

Lea

d T

ime

(hr)

0

500

1000

1500

2000

# E

vent

s

NORTHEAST Winter Storm WarningsLead Time - 2 Year Running Mean

810121416182022

2-Seasons

Lea

d T

ime

(hr)

050010001500200025003000

NORTHEAST Winter Storm WarningsLead Time - 3 Year Running Mean

810121416182022

3-Seasons

Lea

d T

ime

(hr)

05001000150020002500300035004000

KEY ACTIVITIES AND KEY ACTIVITIES AND DEVELOPMENTS DEVELOPMENTS

RELATED TO RELATED TO FORECASTING WINTER FORECASTING WINTER

STORMSSTORMS1993-20041993-2004

NWP ADVANCESNWP ADVANCES

Global ModelGlobal Model 1993 – AVN/MRF at T126/L28 1993 – AVN/MRF at T126/L28

2/day AVN to 126h and 1/day MRF to 240h2/day AVN to 126h and 1/day MRF to 240h 2004 – GFS at T256/L64 to 84 h2004 – GFS at T256/L64 to 84 h

4/day to 384h (T170/L42 84-180h T126/L28 to 386h)4/day to 384h (T170/L42 84-180h T126/L28 to 386h) Numerous improvements to physics, data assimilation etc…Numerous improvements to physics, data assimilation etc…

Medium Range Ensemble Forecast Medium Range Ensemble Forecast System (MREF)System (MREF) ~1997 – 1/day 12 member system~1997 – 1/day 12 member system 2004 – 4/day 48 member system including lag members2004 – 4/day 48 member system including lag members

NGMNGM – Static Since 1993– Static Since 1993

NWP ADVANCESNWP ADVANCES

Eta (NAM)Eta (NAM) 1993 – Early Eta 80 km / 38 levels / 00 and 12Z / 48h1993 – Early Eta 80 km / 38 levels / 00 and 12Z / 48h 1995 – Meso Eta 32 km / 50 levels / 03 and 15Z / 33h1995 – Meso Eta 32 km / 50 levels / 03 and 15Z / 33h 2004 – MesoEta 12 km / 60 levels / 4 x day / 84 hr2004 – MesoEta 12 km / 60 levels / 4 x day / 84 hr Numerous improvements to physics, data assimilation etc…Numerous improvements to physics, data assimilation etc…

Local Area Modeling - Workstation EtaLocal Area Modeling - Workstation Eta

Short Range Ensemble Forecast System (SREF)Short Range Ensemble Forecast System (SREF) ~1999 – 10 members (5 48 km Eta / 5 RSM)~1999 – 10 members (5 48 km Eta / 5 RSM) 2004 – 15 members (10- 32 km Eta / 5- 40 km RSM)2004 – 15 members (10- 32 km Eta / 5- 40 km RSM)

Rapid Update Cycle (RUC)Rapid Update Cycle (RUC) ~1994 – RUC1 – 60 km / 25 levels~1994 – RUC1 – 60 km / 25 levels 1998 – RUC2 – 40 km / 40 levels1998 – RUC2 – 40 km / 40 levels 2004 – RUC2 – 20 km / 50 levels2004 – RUC2 – 20 km / 50 levels

1994-19971994-1997 WSR-88D Network Installation CompletedWSR-88D Network Installation Completed

PCGRIDDS/NAWIPS/GARPPCGRIDDS/NAWIPS/GARP Gridded Model Data into Field OfficesGridded Model Data into Field Offices

NWS Lake Effect Snow StudyNWS Lake Effect Snow Study BUFKITBUFKIT Expansions of Snow Spotter NetworksExpansions of Snow Spotter Networks

Expansion of Collaborative Research Expansion of Collaborative Research Projects (COMET)Projects (COMET) First Real-time local model in NWS ER field office First Real-time local model in NWS ER field office

(MM5 at BUF and BGM)(MM5 at BUF and BGM)

1998-20001998-2000

AWIPS Installations CompletedAWIPS Installations Completed Advanced Workstations w/ Integrated Data SetsAdvanced Workstations w/ Integrated Data Sets

Public Forecast Program TransfersPublic Forecast Program Transfers 1998-99 – BGM/CTP start Winter Warning Program1998-99 – BGM/CTP start Winter Warning Program 1999-00 – AKQ/BTV/CAR/RNK start Winter Warning 1999-00 – AKQ/BTV/CAR/RNK start Winter Warning

ProgramProgram Pros – New ideas/techniques, Smaller Forecast AreasPros – New ideas/techniques, Smaller Forecast Areas Cons – Less experienceCons – Less experience

Active Teletraining Program EstablishedActive Teletraining Program Established Web based training modules Web based training modules

1998-20001998-2000

Regional Workshops/ConferencesRegional Workshops/Conferences Northeast Regional Operational WorkshopNortheast Regional Operational Workshop Southern New England WorkshopSouthern New England Workshop Great Lakes Operational Meteorology WorkshopGreat Lakes Operational Meteorology Workshop Northeast Storm ConferenceNortheast Storm Conference

Continued Expansion of COMET ProjectsContinued Expansion of COMET Projects

CSTAR Collaborative Research Projects CSTAR Collaborative Research Projects FundedFunded North Carolina State Univ. – WFO Raleigh (early North Carolina State Univ. – WFO Raleigh (early

2000)2000) State Univ. of New York at Albany – WFO Albany State Univ. of New York at Albany – WFO Albany

(late 2000)(late 2000)

2001-20042001-2004

Eastern Region Winter Weather Best Eastern Region Winter Weather Best Practices Team Practices Team (2001)(2001)

IFPSIFPS Collaborative Forecast ProcessCollaborative Forecast Process

HPC Winter Weather Experiment (WWE)HPC Winter Weather Experiment (WWE) 2001-02 – 4 WFOs (AKQ/LWX/PHI/CTP)2001-02 – 4 WFOs (AKQ/LWX/PHI/CTP) 2002-03 – All Eastern Region2002-03 – All Eastern Region 2003-04 – ~75% of CONUS2003-04 – ~75% of CONUS 2004-05 – Integrated into routine HPC Operations2004-05 – Integrated into routine HPC Operations

Test new collaborative forecast process among Test new collaborative forecast process among HPC and WFOs for winter storm eventsHPC and WFOs for winter storm events

Evaluate new products from SREFEvaluate new products from SREF

2001-20042001-2004 Implementation of CSTAR Research Implementation of CSTAR Research

ResultsResults U. at Albany ProjectU. at Albany Project

Mesoscale Banding in Winter StormsMesoscale Banding in Winter Storms Precipitation MicrophysicsPrecipitation Microphysics Cold Season Closed LowsCold Season Closed Lows Terrain-Forced Snow StormsTerrain-Forced Snow Storms Impacts of Climate Regimes (ENSO, NAO, etc…)Impacts of Climate Regimes (ENSO, NAO, etc…)

N.C. State ProjectN.C. State Project Cold Air DammingCold Air Damming Coastal FrontsCoastal Fronts Precipitation Type Forecasting MethodologiesPrecipitation Type Forecasting Methodologies

Regional COMET ProjectsRegional COMET Projects BTV/McGill Univ.BTV/McGill Univ. OKX/Stony Brook Univ.OKX/Stony Brook Univ.

2001-20042001-2004

Weather Event Simulator (WES)Weather Event Simulator (WES) – 2001 – 2001 Training Workstation that can simulate real-Training Workstation that can simulate real-

time data flow and forecast processestime data flow and forecast processes

AWIPS Archive ServerAWIPS Archive Server - 2002- 2002 Local Capability to archive full AWIPS data sets Local Capability to archive full AWIPS data sets

for WES playbackfor WES playback

Expansion of local office training Expansion of local office training activities and workshopsactivities and workshops

2001-20042001-2004

Coming together of several Coming together of several activities and developments:activities and developments: Applied Research and Technique Applied Research and Technique

DevelopmentDevelopment TechnologiesTechnologies TrainingTraining Operational Application and ProceduresOperational Application and Procedures

Mesoscale Band Project Mesoscale Band Project TimelineTimeline

RELATIONSHIP BETWEEN RELATIONSHIP BETWEEN EVENT COUNTS AND EVENT COUNTS AND

WARNING PERFORMANCEWARNING PERFORMANCE

IMPACTS OF CLIMATE IMPACTS OF CLIMATE REGIMES?REGIMES?

ACTIVE SEASONS IN ACTIVE SEASONS IN NORTHEASTNORTHEAST

>1200 Events>1200 EventsSeasoSeaso

nn# of # of

EventsEventsPODPOD FARFAR Lead Lead

TimeTime

93-94*93-94* 14161416 .923.923 .230.230 11.811.8

95-9695-96 17131713 .842.842 .246.246 14.614.6

00-0100-01 12201220 .913.913 .296.296 18.018.0

02-0302-03 13831383 .916.916 .277.277 16.216.2

MeanMean .899.899 .262.262 15.215.2

MODERATE SEASONS IN MODERATE SEASONS IN NORTHEASTNORTHEAST

800-1200 Events800-1200 EventsSeasoSeaso

nn# of # of

EventsEventsPODPOD FARFAR Lead Lead

TimeTime

98-9998-99 10501050 .890.890 .314.314 13.613.6

99-0099-00 10161016 .902.902 .201.201 12.712.7

03-0403-04 10691069 .920.920 .321.321 20.320.3

MeanMean .904.904 .279.279 15.515.5

LAMELAME SEASONS IN NORTHEASTSEASONS IN NORTHEAST <800 Events<800 Events

SeasoSeasonn

# of # of EventsEvents

PODPOD FARFAR Lead Lead TimeTime

94-9594-95 561561 .772.772 .467.467 13.613.6

96-9796-97 654654 .855.855 .538.538 11.311.3

97-9897-98 788788 .803.803 .489.489 12.812.8

01-0201-02 470470 .840.840 .402.402 11.711.7

MeanMean .818.818 .474.474 12.312.3

MONTHLY NAO/PNA vs. MONTHLY NAO/PNA vs. EventsEvents

Cumulative Winter (Dec-Jan-Feb-Mar) Monthly Mean NAO shows Cumulative Winter (Dec-Jan-Feb-Mar) Monthly Mean NAO shows some correlation to number of winter storm eventssome correlation to number of winter storm events

Cumulative Winter Monthly Mean PNA shows little or no correlationCumulative Winter Monthly Mean PNA shows little or no correlation

Shorter term (daily/weekly) index values likely more importantShorter term (daily/weekly) index values likely more important

SUMMARYSUMMARY

Winter storm warning performance Winter storm warning performance appears to be improving across the appears to be improving across the NortheastNortheast Greatest improvements in Lead TimeGreatest improvements in Lead Time

Lead Time improvements are not a result of Lead Time improvements are not a result of improved PODimproved POD

No increase in false alarms notedNo increase in false alarms noted

Event totals impact warning performanceEvent totals impact warning performance More events – better performanceMore events – better performance Most impact on False Alarms and Lead TimeMost impact on False Alarms and Lead Time

SUMMARYSUMMARY

Performance improvements appear Performance improvements appear

to be related to an evolving “end-to be related to an evolving “end-

to-end” collaborative process of:to-end” collaborative process of:

Discovery and SharingDiscovery and Sharing

Demonstration of Added Value (Operational Utility)Demonstration of Added Value (Operational Utility)

Operational ImplementationOperational Implementation

Training ActivitiesTraining Activities

Periodic Review and RefinementPeriodic Review and Refinement