a multi-parameter approach to lightning prediction gail hartfield noaa/nws raleigh, north carolina...
TRANSCRIPT
A Multi-Parameter Approach to Lightning Prediction
Gail HartfieldNOAA/NWS
Raleigh, North Carolina
2011 American Meteorological Society Annual Meeting, Seattle, WA
Much value can be gained from a skillful forecast of lightning activity.
Accurate lightning prediction remains challenging Process still poorly understood Incomplete metrics (only CGs
are detected over much of U.S.)
Improved near-term modeling and observations are promising
Predictions could have enormous benefits for the weather enterprise and related/dependent groups Aviation Utilities Recreation/outdoor activities
U.S. Lightning Deaths, 2000-2009
Map courtesy of Ronald Holle, Vaisala, Inc., and B.Curran and M. Bragaw,and NASA MSFC Lightning Imaging Sensor (LIS) Science Team
Several groups are working toward lightning prediction in different ways.
Research into improved understanding (e.g. LMA applications) (e.g. SPoRT/NASA, NWS, NSSL)
Flow regime analysis and climatologies (e.g. FSU, TAMU)
“First strike” radar-based detection (NSSL, NCSU)
Occurrence of any lightning (e.g. Patrick AFB/KSC)
Lightning threat parameters in convection-allowing models (e.g. NSSL, SPoRT/NASA)
Focus on occurrence of excessive lightning (high density, or HD, lightning) with an ingredients-based method (e.g. NWS Raleigh)
Strike density climatology
WFO Raleigh’s Analysis and Prediction of the Potential for Excessive Lightning (APPEL) project:
Comprised of a forecast component and an analysis component
24 Hour CG Lighting Strikes in the RAH CWA
0
2000
4000
6000
8000
10000
12000
14000
16000
7/1
5/2
00
8
8/1
5/2
00
8
9/1
5/2
00
8
10
/15
/20
08
11
/15
/20
08
12
/15
/20
08
1/1
5/2
00
9
2/1
5/2
00
9
3/1
5/2
00
9
4/1
5/2
00
9
5/1
5/2
00
9
6/1
5/2
00
9
7/1
5/2
00
9
Date
CG
Lig
hti
ng
Str
ikes
(Experimental forecasts based on applicationof peer-reviewed lightning research)
(Climatology + case studies of events with highdensity lightning)
Our goal: to produce a skillful 3-24 hr outlook of lightning activity.
Approach: a multi-parameter, ingredients-based checklist (thresholds based on past research and local cases) with forecaster input
Why a checklist? Can use preferred model Not a “black box” Forecaster can apply experience and
weigh parameters accordingly Focus: catch the “big” lightning days,
and include this info in the morning Hazardous Weather Outlook
Have completed two seasons (May-Sept) of forecasting
Checklist components and thresholds were based on past studies of environmental conditions preceding excessive lightning events.
Moistureand
instability
Instability and CAPE “shape”
Multi-parameter
indices fromSPC
Precip. water
Subjective assessmentsof forcing and
moisture
Helps assess two critical components for lightning production:
• Available moisture (the presence of graupel is essential for electrification)
• Instability (especially aloft) (the greater the buoyancy, the more vigorous the updrafts)
Forecasters have several tools at their disposal to help fill out the checklist:
AWIPSprocedures
BUFKIT forecast soundings
Local WRF model
Forecasters have several tools at their disposal to help fill out the checklist:
SPC NAM-based“perfect prog”forecast
New this year: NSSL WRF-basedtotal lightning threat
SPC SREF-based calibratedprobabilities of >100 CGs
Lightning activity forecast results for 2009: Enhanced lightning threat added to the Hazardous
Weather Outlook (HWO) on 24 of 153 days Several noted the extreme nature of expected lightning (“…
will be nearly continuous…as much as one strike every few seconds…”)
Enhanced wording included for 3 of the 4 greatest lightning density days
Local media has begun to incorporate lightning prediction into broadcasts
“IN ADDITION...SUFFICIENT INSTABILITY IS PRESENT AT MID AND UPPER LEVELS OF THE ATMOSPHERE TO SUPPORT AN UNUSUALLY LARGE AMOUNT OF LIGHTNING THIS AFTERNOON AND EVENING.”
“STRONG THUNDERSTORMS MAY PRODUCE FREQUENT LIGHTNING...IN SOME CASES ONE STRIKE EVERY 10 SECONDS.”
BUT… there were a few problems: Forecaster experience and confidence varied Inclusion in HWO somewhat inconsistent by forecaster and situation Extra duties took time
Challenges remain with this technique. Is it ideal to apply mesoscale characteristics to the storm
scale? How do we weigh dynamic and thermodynamic
contributions to ascent? Both parameterized and convection-allowing models are
used, and these must be evaluated properly Can be inconsistency among forecasters
Our latest efforts: Complete detailed areal climatology (to help define a
“significant” day) Case studies and composites (to improve pattern
recognition capabilities and increase understanding of HD lightning events)
Initial operational evaluation of NSSL WRF-based lightning threat
OUCH!
2002-2010 data analysis reveals year-to-year variations. (Large scale flow pattern differences?)
For 2002-2010:• Average lightning days per year: 100
• Lightning on top 10 days, on average, accounts for 49.9% of yearly total lightning
• Most electrically active systems were multicell clusters and dynamically driven convective lines
Area of study
Data source: NLDN
= RAH CWA
= Verification area
Lightning analysis for June 2002-September 2010 reveals yearly maxima from late June through late August.
Data analysis courtesy of Jeremy Gilchrist and Whitney Rushing, NCSU
20032004
2005
2006
2007 20082009
2002
2010
Area of study
= RAH CWA
= Verification area
Data source: NLDN
Lightning analysis for June 2002-September 2010 reveals yearly maxima from late June through late August.
Area of study
Data source: NLDN
28 July 2005 (most all-time): 39,754 strikes in central NC
CGs, 1200 UTC 28 Jul to 1200 UTC 29 Jul
Comparing top HD lightning days with average days in 2010:
High lightning activity days (top 10%)
Average lightning activity days
MLCAPE 2550 1405
MUCAPE 3400 2110
CAPE (-10°C to -30°C) 620 390
Normalized CAPE 0.21 0.13
Precipitable water 2.22 1.83
∆ PW (T-6hrs to T0hrs) +0.18 +0.10
Suggests high potential for strong instability and
vigorous updrafts
Gauges potential for “fat” CAPE and strong instability in the mixed
phase layer aloft
Indicates presence of sufficient graupel and
graupel flux
Parameters from RUC-based SPC mesoanalyses, within 100 km of convection, within 2 hours of onset
High lightning activity days (top 10%)
Average lightning activity days
MLCAPE 2550 1405
MUCAPE 3400 2110
CAPE (-10°C to -30°C) 620 390
Normalized CAPE 0.21 0.13
Precipitable water 2.22 1.83
∆ PW (T-6hrs to T0hrs) +0.18 +0.10
Findings consistent with past research Guidelines for potential high-density events:
MUCAPE > 3000 J/kg MLCAPE > 2000 J/kg PW > 150% of normal & positive PW flux
Blended TPW products have been helpful
Findings consistent with past research Guidelines for potential high-density events:
MUCAPE > 3000 J/kg MLCAPE > 2000 J/kg PW > 150% of normal & positive PW flux
Blended TPW products have been helpful
PW climatology courtesy of NWS Rapid City, SD
Typical PW for 2010 high-density lightning days
Blended TPW imagery, available on the web and onNWS AWIPS systems
Composites of precipitable water and surface lifted index, 2010 top 10 high-density (l.) and 10 average (median, r.) lightning days
PW=45-55 mm PW=35-45 mm
LI= -3 to -3.5 C LI= -1 to -2 C
Average lightning daysGreatest lightning days
5 August 2010: 10,642 CG strikes in central NC
CGs, 1200 UTC 5 Aug to 1200 UTC 6 Aug
2 people struck by lightning in Raleigh(struckbylightning.org)
5-6 August 2010
KRAX 88D reflectivity
NSSL WRF lightning threat
Hourly CG strikes
05/0300 UTC 05/2000 UTC 06/0000 UTC
SPC WRF, 1 km refl, 0400 UTC (16h fcst)
KRAX 88D, reflectivity, 0334 UTC
Local WRF-NMM, 1 km refl, 0400 UTC (16h fcst)
NSSL WRF lightning threat, 0400 UTC (28h fcst)
25 July, 20109,497 CG strikes
CGs, 25 Jul – 26 Jul, 1200-1200 UTC
Conclusions & future plans What we know:
NSE analysis for HD lightning days supports past research and measurements Confirms importance of graupel flux and strong updrafts for lightning production
Skillful lightning prediction can be done WRF-, NAM-, and SREF-based output very helpful
Augment NSE parameters Forecaster intervention may improve upon purely automated methods Working together and sharing findings are critical for success
What’s coming up: Resumption of forecasts in May 2011 Continued operational evaluation of WRF-based lightning threat fields Creation of an experimental “ensemble” parameter for each model Verification of automated methods vs. human-based method
NWS Raleigh Lightning Team: Jonathan Blaes (NWS), Morgan Brooks (NWS), Jeremy Gilchrist (NCSU), Whitney Rushing (NCSU), Gail Hartfield (NWS – [email protected])
EXTRA SLIDES
Year Top Quartile Top 10 sum Top 25 sum# Days over
2K Biggest day # Date Yrly Total LtgDays Top10/Total
2002 3939 78759 121782 21 20337 6-Jul-02 132482
2003 4023 123670 212032 45 20851 17-Jul-03 302412 101 40.9%
2004 2768 101143 161303 36 17332 24-May-04 221547 107 45.7%
2005 2012 160150 214113 25 39754 29-Jul-05 248166 96 52.0%
2006 3728 142874 223919 30 27795 31-Aug-06 274290 91 52.0%
2007 2015 88085 133176 25 13376 11-Jul-07 162710 97 54.1%
2008 2210 88660 144709 27 11734 9-Jul-08 184261 103 48.1%
2009 944 79813 112981 18 14713 10-Jun-09 133547 111 59.8%
2010 2530 80044 139521 29 10642 6-Aug-10 172060 90 46.5%
Average
28 100 49.9%
Yearly CG strike statisticsWFO RAH CWA
Average lightning days (r., compared to top days, l.) had lower heights aloft and more ridging over the Southeast.
500 mb mean heights/anomalies,top 10 days of 2010
500 mb mean heights/anomalies,average days of 2010
500 mb composites of top 2009 & 2010 days indicates a mid level trough to our northwest (but signal is weak).
500 mb mean heights/anomalies,top 10 days of 2010 (80044 CG strikes)
500 mb mean heights/anomalies,top 10 days of 2009 (79813 CG strikes)
CGs similar for both years, but 2010 was hotter (more CAPE?)and 2009 cases appear more dynamic (compensating for lower CAPE?)
CGs similar for both years, but 2010 was hotter (more CAPE?)and 2009 cases appear more dynamic (compensating for lower CAPE?)
500 mb composite of top days indicates a mean weak mid level trough to our west.
500 mb mean heights,top 10 days of 2010
500 mb mean heights,top 10 days 2002-2010
Partial bibliography for the Analysis and Prediction of the Potential for Excessive Lightning (APPEL) project
Keller, D. L., 2004: Forecasting cloud-to-ground lightning data with AFWA-MM5 model data using the "Bolt of Lightning Technique" (BOLT) algorithm. Preprints, 22nd Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., CD preprints.
Hondl, K. and M. Eilts, 1994: Doppler radar signatures of developing thunderstorms and their potential to indicate the onset of cloud-to-ground lightning. Mon. Wea. Rev., 122, 1818–1836.
Wolf, P., 2006: Anticipating the Initiation, Cessation, and Frequency of Cloud-to-Ground Lightning Utilizing WSR-88D Reflectivity Data (local study, WFO JAX). Shafer, P.E., 2005: Developing Gridded Forecast Guidance for Warm Season Lightning over Florida Using the Perfect Prognosis Method and the Weather Research & Forecast
Model (doctoral prospectus). Bothwell, P., 2005: Development of an operational statistical scheme to predict the location and intensity of lightning. Preprints, Amer. Meteor. Soc. Annual Meeting, 2005. Deierling, W. and W. A. Petersen, J. Latham, S. M. Ellis, H. J. Christian Jr. and J. Walters, 2006: Total lightning frequency in relation to ice masses and ice mass flux estimates.
Preprints, Amer. Meteor. Soc. Annual Meeting, 2006. Shafer, P.E. and H. Fuelberg, 2005: A statistical procedure to forecast the daily amount of warm season lightning in south Florida. Preprints, Amer. Meteor. Soc. Annual
Meeting, 2005. Cope, A., 2006: Toward better use of lightning data in operational forecasting. Preprints, Amer. Meteor. Soc. Annual Meeting, 2006. Blanchard, D.O., 1998: Assessing the vertical distribution of convective available potential energy (CAPE). Wea. Forecasting, Sep. 1998. Jayaratne, R. and E. Kuleshov, 2006: The relationship between lightning activity and surface wet bulb temperature and its variation with latitude in Australia. Meteorology and
Atmospheric Physics, 91, pp.17-24. Petersen, W. A. and S. Rutledge, 2001: Regional variability in tropical convection: observations from TRMM. Journal of Climate, Sep. 2001. Petersen, W. A., 1997: Multi-Scale Process Studies in the Tropics: Results from Lightning Observations. Doctoral thesis, Colo. St. Univ., 1997. Van Den Broeke, M. and D. Schultz, R. Johns, J. Evans, and J. Hales, 2004: Cloud-to-Ground Lightning Production in Strongly Forced, Low-Instability Convective Lines Associated
with Damaging Wind. Wea. Forecasting, Aug. 2005. MacGorman, D.R., and W.D. Rust, 1998: The Electrical Nature of Storms. Oxford University Press, 422 pp. Bright, D.R. and M. S. Wandishin, R. E. Jewell, and S. J. Weiss, 2005: A physically based parameter for lightning prediction and its calibration in ensemble forecasts. Preprints,
Amer. Meteor. Soc. Annual Meeting, 2005. Kehrer, K. and B. Graf and W. Roeder, 2008: Global Positioning System (GPS) precipitable water in forecasting lighting at Spaceport Canaveral. Wea. Forecasting, Apr. 2008. Lambert, W. and D. Sharp, S. Spratt, and M. Volkmer, 2006: Using cloud-to-ground lightning climatologies to initialize gridded lightning threat forecasts for east central Florida.
Preprints, Amer. Meteor. Soc. Annual Meeting, 2006. Murphy, M. and C. E. Konrad, 2005: Spatial and temporal patterns of thunderstorm events that produce cloud-to-ground lightning in the interior southeastern United States.
Monthly Weather Review, 133, 1417-1430. Williams, E. and V. Mushtak, D. Rosenfeld, S. Goodman, and D. Boccippio, 2005: Thermodynamic conditions favorable to superlative thunderstorm updraft, mixed phase
microphysics, and lightning flash rate. Atmos. Research, 76, 2005. McCaul, Jr., E. and K. LaCasse, S. Goodman, and D. Cecil, 2008: Use of high-resolution WRF simulations to forecast lightning threat. Preprints, Amer. Meteor. Soc. Annual
Meeting, 2008. Deierling, W. and W. A. Petersen, J. Latham, S. M. Ellis, and H. J. Christian Jr., 2005: Towards the relationship between total lightning activity and downward as well as upward
ice mass fluxes in thunderstorms. Preprints, Amer. Meteor. Soc. Annual Meeting, 2005. Mazany, R. and S. Businger, S. Gutman, and W. Roeder, 2002: A lightning prediction index that utilizes SPS integrated precipitable water vapor. Wea. Forecasting, Oct. 2002. Livingston, E. and J Nielsen-Gammon and R. Orville, 1996: A climatology, synoptic assessment, and thermodynamic evaluation for cloud-to-ground lightning in Georgia: a study
for the 1996 Summer Olympics. Bull. Amer. Meteor. Soc., 77, Jul. 1996. Burrows, W. and C. Price and L. Wilson, 2004: 1 to 2 day prediction of the probability of lightning occurrence over Canada and the northern United States in the warm season.
Canadian Meteorological and Oceanographic Society Annual Meeting, 2004.