1 Storm Prediction Center Use of GPS IPW Data as an Input to Severe Weather Forecasting Steven Weiss [email protected] NOAA GPS/GNSS Workshop October

1

Storm Prediction Center Use of GPS IPW Data as an Input to Severe Weather Forecasting

Steven [email protected]

NOAA GPS/GNSS Workshop

October 24, 2007Boulder, CO

2

Outline

• Brief Overview of Storm Prediction Center• Severe Thunderstorm Forecasting

– Importance of observational data for short-term hazardous weather prediction

• Relationship between moisture and deep convection

– Sources of observational data in operational forecasting

• Sampling and resolution issues

• Use of GPS-IPW at Storm Prediction Center– Several case examples

• Conclusions

3

4

STORM PREDICTION CENTER

MISSION STATEMENT

The Storm Prediction Center (SPC) exists

solely to protect life and property of the American people

through the issuance of timely, accurate watch and forecast products

dealing with tornadoes, wildfires and other hazardous mesoscale weather

phenomena.

MISSION STATEMENT

The Storm Prediction Center (SPC) exists solely to Protect Life and PropertyProtect Life and Property of the American people through the issuance of timely, accurate watch and forecast products dealing with hazardous mesoscale weather phenomena.

5

• Hail, Wind, Tornadoes

• Excessive rainfall

• Fire weather

• Winter weather

STORM PREDICTION CENTER HAZARDOUS PHENOMENA

6

• TOR / SVR TSTM WATCHES (~1000 / yr)– Valid for 6-8 hours

• WATCH STATUS REPORTS (~5000 / yr)– Valid 0-1 hour

• CONVECTIVE OUTLOOKS (~3200 / yr)– Day 1; Day 2; Day 3; Days 4-8

• MESOSCALE DISCUSSIONS (~2000 / yr)– Valid for 1-3 hours– Severe Potential Preceding Watch Issuance– Analysis of Severe Weather in Active Watches– Heavy Rainfall – Hazardous Winter Weather

• FIRE WEATHER OUTLOOK (~1200 / yr)– Day 1; Day 2; Days 3-8

• FORECASTS ISSUED IN DETERMINISTIC AND PROBABILISTIC FORMATS

– Text and Graphical Products

SPC Forecast Products

75% of all SPC products are valid for < 24h period- Observational data are critical for short-term forecasting

7

Severe Weather Outlooks

•Two Outlook Types

•Categorical•Slight Risk

•Moderate Risk

•High Risk

•Probabilistic•Tornadoes

•Hail

•Convective Winds

8

Example of High Risk Outlook DayDay 1 Outlook 7 April 2006

Categorical Risk Tornado Probability

Hail Probability Wind Probability

9

What Happened on 7 April 2006?

10

Mesoscale Convective Discussions (MD)

• Goal is to issue pre-watch MDs 1 to 3 hours prior to a Severe Thunderstorm or Tornado watch issuance.

- Define area(s) of concern - State expected watch type - Provide meteorological reasoning – most important

• Also issued ~ every 2 hours for each active watch to provide diagnostic/short-term forecast information (when WFOs are busy with warning activities)

11

Example of Pre-Watch MDMESOSCALE DISCUSSION 2112 NWS STORM PREDICTION CENTER NORMAN OK0159 PM CDT THU OCT 18 2007

AREAS AFFECTED...LOWER MI CONCERNING...SEVERE POTENTIAL...TORNADO WATCH LIKELY VALID 181859Z - 182000Z

A FEW STORMS COULD DEVELOP OVER THE NEXT COUPLE HOURS. ISOLATED TORNADOES AND HAIL WOULD BE LIKELY. VISIBLE IMAGERY SHOWS CLEARING HAS TAKEN PLACE OVER MUCH OF LOWER MI. HORIZONTAL ROLLS BENEATH CAPPING INVERSION HAVE CHANGED CHARACTER OVER WRN MI SUGGESTING CAPPING IS BEING REMOVED BY COOLING ALOFT AND SURFACE HEATING. ENHANCED CONVERGENCE ALONG LAKE BREEZE LOOKS TO BE A FOCUS FOR DEVELOPMENT OVER THE NEXT FEW HOURS...POSSIBLY EXTENDING AS FAR S AS NRN IN WHERE CU FIELD IS ALSO INCREASING AND

DEWPOINTS ARE RISING. CONTINUED HEATING AS WELL AS COOLING ALOFT SHOULD ALLOW A FEW STORMS TO DEVELOP. SHEAR PROFILES ARE VERY STRONG AND ROTATING UPDRAFTS ARE LIKELY...ALONG WITH ISOLATED TORNADOES.

..JEWELL.. 10/18/2007 ATTN...WFO...DTX...APX...IWX...GRR...LOT...

12

SPC Convective Watch GoalsSPC Convective Watch Goals

• WATCHES ATTEMPT TO CAPTURE:WATCHES ATTEMPT TO CAPTURE:– ALL SIGNIFICANT SEVERE EVENTS:ALL SIGNIFICANT SEVERE EVENTS:

• 2” or larger hail2” or larger hail

• 65+ kt wind gusts65+ kt wind gusts

• EF2+ TornadoesEF2+ Tornadoes

– MULTIPLE SEVERE EVENTS FROM ORGANIZED MULTIPLE SEVERE EVENTS FROM ORGANIZED CONVECTIONCONVECTION

• SupercellsSupercells

• Squall linesSquall lines

• Mesoscale Convective Systems (MCSs)Mesoscale Convective Systems (MCSs)

– ISSUE 1-2 HRS BEFORE ONSET OF SEVEREISSUE 1-2 HRS BEFORE ONSET OF SEVERE• Allows WFOs, Emergency Managers, Media, etc. time to Allows WFOs, Emergency Managers, Media, etc. time to

implement hazardous weather contingency plansimplement hazardous weather contingency plans– Deploy storm spottersDeploy storm spotters– Increase staffingIncrease staffing– Contact critical customers (e.g., hospitals, utility companies)Contact critical customers (e.g., hospitals, utility companies)

13

URGENT - IMMEDIATE BROADCAST REQUESTED TORNADO WATCH NUMBER 44 NWS STORM PREDICTION CENTER NORMAN OK 730 AM CST THU MAR 1 2007

THE NWS STORM PREDICTION CENTER HAS ISSUED A TORNADO WATCH FOR PORTIONS OF MUCH OF ALABAMA PARTS OF EASTERN ARKANSAS WESTERN FLORIDA PANHANDLE SOUTHEASTERN LOUISIANA MISSISSIPPI WESTERN AND MIDDLE TENNESSEE COASTAL WATERS

EFFECTIVE THIS THURSDAY MORNING AND EVENING FROM 730 AM UNTIL 600 PM CST.

...THIS IS A PARTICULARLY DANGEROUS SITUATION...

DESTRUCTIVE TORNADOES...LARGE HAIL TO 2.5 INCHES IN DIAMETER... THUNDERSTORM WIND GUSTS TO 80 MPH...AND DANGEROUS LIGHTNING ARE POSSIBLE IN THESE AREAS. THE TORNADO WATCH AREA

Example of SPC Tornado Watch

14

Watch Hazard Probabilities

Tornado (2+ reports): High (>95%)

EF2+ Tornado: High (70%)

Wind (10+ reports): High (80%)

75 mph Wind: High (70%)

Hail (10+ reports): High (70%)

2”+ Hail: Mod (60%)

Hail & Wind (6+ reports): High (>95%)

Watch Counties -- Watch Probabilities

Note: The Aviation Watch (SAW) parallelogram product is an approximation to the watch area. The actual watch is depicted by the shaded county areas.

15

Severe Thunderstorm Severe Thunderstorm ForecastingForecasting

16

Modified Forecast Funnel• SPC focuses on relationship between synoptic - mesoscale environment and subsequent thunderstorm development and evolution

• Must maintain awareness of mesoscale - synoptic scale interactions

• Severe weather events occur on scales smaller than standard observational data (and traditional model output)

• The real atmosphere is more important than a model atmosphere

• How well do we sample environment?

17

Physical Processes for Deep Convection

Ingredients for deep moist convection:

1) Moisture 2) Instability 3) Lift

This sounds straightforward, but…

18


• Moisture

• Typically focus on PBL moisture (dew point or mixing ratio)

• But moisture sources above PBL can be associated with elevated convection

• Instability

• Temperature lapse rates and moisture are key contributors (juxtaposition of steeper lapse rates above moist layer)

• Sufficient CAPE within hydrometeor mixed phase region

• CAPE twice as sensitive to lifted parcel moisture than to parcel temperature

19


• Moisture

• A major limitation in convective forecasting is the lack of high resolution 4D water vapor data

• Observing system does not sufficiently resolve details needed to properly “initialize” forecasters, researchers, and NWP models

• NWP models reflect strengths / weaknesses of each model as well as observational limitations (physics and IC uncertainty)

• Lack of necessary detail in mapping water vapor has been identified by USWRP as major obstacle to improved QPF and severe weather forecasting

• New sources of accurate and reliable observational data are critical to improve assessment of environment and convective forecasting

20

Some Observational Sources for Atmospheric Water Vapor Data

21

Common Operational Observing Systems

• All observing systems have strengths and limitations– Radiosondes

• High vertical resolution, low time and horizontal resolution

– Surface METAR• High horizontal and time resolution, no vertical information

– Wind profilers and NEXRAD VAD winds• High vertical and time resolution, moderate horizontal res.• No thermodynamic data

– Satellite retrievals (winds and thermodynamic)• Variable horizontal and time resolution, low vertical resolution• Sounder impacted by clouds; microwave best over oceans

– ACARS/TAMDAR aircraft data• Variable time/space resolution; limited water vapor data

22

GPS Integrated Precip. Water (IPW)

• GPS IPW sampling characteristics– High time resolution., mod. horizontal. res., low vertical res.

• GPS IPW data have unique attributes that enhance ability of SPC forecasters to monitor moisture distribution and transport– All weather sensing (not affected by clouds and precipitation)

– CONUS domain

– Accurate (observational uncertainty 1-2 mm)

– Continuous operating capability

• Lack of vertical moisture profiles compensated by knowledge that majority of water vapor is located in lowest 2-3 km AGL– Changes in GPS IPW largely reflect changes in low level moisture field

23

GPS Integrated Precip. Water (IPW)

Schematic of vertical distribution of water vapor and temperature

• Water vapor mixing ratio decreases rapidly with height in the atmosphere

• ~50% of IPW in the lowest 1.5 km

• ~75% of IPW in lowest 3 km

• Less than 6% of IPW above 5 km

Source: Water Vapor in Climate System Special Report, AGU (1995)

24

GPS IPW at the Storm Prediction Center

25

GPS IPW at the Storm Prediction Center

• SPC forecasters directly access GPS IPW data three ways– Hourly station plots in N-AWIPS workstation displays– GSD web page (http://gpsmet.fsl.noaa.gov/jsp/index.jsp)

– Six-hourly internet graphics from CIRA - CSU web page (http://amsu.cira.colostate.edu/gpstpw/)

• GPS IPW data also are part of EMC data assimilation for the RUC and NAM models– RUC model analyses and forecasts improved by GPS IPW data– At SPC, RUC analyses and 1-hr forecasts combined with hourly

surface data to create SPC 3D Mesoscale Analyses (http://www.spc.noaa.gov/exper/mesoanalysis/)

– SPC Mesoscale Analyses used heavily by SPC and NWS field forecasters during high impact weather events

26

GPS IPW in SPC N-AWIPS Workstations

Data received via link from NCEP/HPC

27

Case Example – 19 Sept 2007

SPC Day 1 Convective Outlook Issued 13 UTC

28

N-AWIPS Display 13 UTC 19 Sept 2007 GPS IPW Plot and Analysis

29


30


31


32


33


34


35


36


37


38


39

Grenada CO GPS IPW Trends 00 UTC 16 Sept-06 UTC 20 Sept

16-00z

40

Blended GPS TPW on CIRA - CSU Web Page

(http://amsu.cira.colostate.edu/gpstpw/)

• GPS over Land and AMSU-SSM/I over Oceans

• Output at 6-hr intervals during past 72 hrs

• Animated Loops and Anomaly Charts

(Case Example: Courtesy Sheldon Kusselson (NESDIS/SAB)

41

CIRA - CSU Blended TPW

AMSU and SSM/I Over Oceans Only Now Include GPS over CONUS

Addition of GPS IPW provides MUCH better support for CONUS forecasters

42

Tornado Outbreak 28 March 2007

Oklahoma Panhandle

Silverton, TX

SPC Severe Reports (preliminary)

red - tornado green - hail blue - wind

43

Blended TPW 18 UTC 28 Mar 2007

Blended TPW Blended TPW Anomaly

44

Blended TPW 00 UTC 29 Mar 2007 Note increase in anomaly magnitudes and moist/dry gradient

Blended TPW Blended TPW Anomaly

MOIST

DRY

ABOVE

BELOW

45

Example of Current Watch MDMESOSCALE DISCUSSION 2095NWS STORM PREDICTION CENTER NORMAN OK0236 PM CDT WED OCT 17 2007

AREAS AFFECTED...ERN TX...LA...SRN ARCONCERNING...TORNADO WATCH 710... VALID 171936Z - 172130Z

THE SEVERE WEATHER THREAT FOR TORNADO WATCH 710CONTINUES. THE THREAT FOR TORNADOES CONTINUES WITHIN WW 710. LOW LEVEL WARM/MOIST ADVECTION CONTINUES WITH GPS WATER VAPOR SENSORS CURRENTLY INDICATING 2.00-2.32 INCHES FROM LA INTO SRN AR.

MEANWHILE...SATELLITE IMAGERY INDICATES LOW LEVEL VORTICITY CENTER MOVING NWD ALONG THE SABINE RIVER...ON CYCLONIC SIDE OF 700 MB JETLET OVER THE LOWER MS VALLEY. THIS FEATURE IS HELPING TO FOCUS CONVECTION AND WILL CONTINUE NWD INTO AR. SHEAR PROFILES ARE ALREADY FAVORABLE FOR ROTATING UPDRAFTS AND TORNADOES...ALTHOUGH RELATIVELY WARM PROFILES AND WEAK LAPSE RATES HAVE ACTED TO SLOW DEVELOPMENT. SHEAR PROFILES WILL LIKELY INCREASE WITH TIME AS UPPER TROUGH CONTINUES TO DEVELOP NEWD. THIS WILL MAINTAIN OR INCREASE TORNADO THREAT FARTHER N INTO AR...AND POSSIBLY SERN MO...WHERE AN ADDITIONAL WW MAY BE REQUIRED LATER TODAY INTO THIS EVENING.

..JEWELL.. 10/17/2007 ATTN...WFO...JAN...LIX...LZK...LCH...SHV...HGX...FWD...

46

Example of Heavy Rain MDMESOSCALE DISCUSSION 1949NWS STORM PREDICTION CENTER NORMAN OK0114 AM CDT SAT SEP 08 2007

AREAS AFFECTED...PORTIONS NRN/WRN OK...E-CENTRAL/NERNPANHANDLE OF TX. CONCERNING...HEAVY RAINFALLVALID 080614Z - 080915Z

BAND OF TSTMS HAS BEEN INCREASING IN COVERAGE FOR ACOUPLE HOURS...FROM CHEROKEE-GARFIELD-ROGER MILLSCOUNTIES OK INTO GRAY COUNTY TX. THIS ACTIVITY SHOULDPERSIST WITH MINOR FLUCTUATIONS IN ITS MESOSCALEGEOMETRY...PERHAPS INCLUDING SLGT NET SWD DRIFT...THROUGH 9Z. RAIN RATES 1-2 INCHES/HOUR WILL BE COMMON...LOCALLY AROUND 2.5 INCHES/HOUR IN HEAVIEST CORES. WITHLITTLE MOVEMENT EXPECTED FOR ASSOCIATED MAIN ZONE OFLOW LEVEL LIFT...PERSISTENT HEAVY RAIN EVENT WITHTRAINING/MERGING OF CELLS WILL ENHANCE HEAVY RAINHAZARD.

…………………………..

RUC FCST SOUNDINGS AND GPS DATA INDICATE PW COMMONLY

AROUND 2 INCHES ACROSS THIS CORRIDOR...AND AS MUCH AS2.3 INCHES INVOF TUL...WITH 500-1500 J/KG MUCAPE INSUPPORT OF TSTM POTENTIAL.

..EDWARDS.. 09/08/2007 ATTN...WFO...TSA...OUN...AMA...

47

Conclusions

• SPC forecasters depend heavily on high quality observational data that provide increased temporal or spatial measurements of atmospheric structure to support our severe storm forecast mission– 4D evolution of water vapor is critical for thunderstorm prediction

• GPS-IPW provides reliable, timely, and accurate measurements of integrated water vapor over land– Most IPW resides below 3 km (proxy for low-level moisture)– Availability of GPS IPW enhances forecaster situational

awareness• Particularly important during dynamic, rapidly changing situations

• GPS IPW improves RUC model analyses/forecasts– RUC is a key input to SPC Mesoscale Analysis fields

Documents

1 Storm Prediction Center Use of GPS IPW Data as an Input to Severe Weather Forecasting Steven Weiss [email protected] NOAA GPS/GNSS Workshop October