Goodman slides

Slide1: 

Use of High-Resolution WRF Simulations to Forecast Lightning Threat Photo, David Blankenship Guntersville, Alabama 1Steve Goodman, 2W. Eugene (Bill) McCaul, Jr., 3Kate LaCasse 1NASA Marshall Space Flight Center, Earth Science Office 2Universities Space Research Association 3University of Alabama in Huntsville Huntsville, Alabama, USA LIS11 International Workshop 11-13 September, 2006 Huntsville Alabama

Outline of Presentation: 

Outline of Presentation Motivation and Background Science and Technology Infusion Nowcasting-Warning Decision Making Weather Research and Forecast Model (WRF) Forecasts of Thunderstorm Initiation, End Forecasts of Lightning Threat, Rate Current Plans and the Way Forward

Science Questions and Focus Areas: 

Variability Forcing Response Consequence Prediction Precipitation, evaporation & cycling of water changing? Global ocean circulation varying? Global ecosystems changing? Atmospheric composition changing? Ice cover mass changing? Earth surface transformation? Atmospheric constituents & solar radiation on climate? Changes in land cover & land use? Motions of the Earth & Earth’s interior? Clouds & surface hydrological processes on climate? Ecosystems, land cover & biogeochemical cycles? Changes in global ocean circulation? Atmospheric trace constituents responses? Sea level affected by Earth system change? Regional air quality impacts? Weather variation related to climate variation? Consequences of land cover & land use change? Coastal region impacts? Weather forecasting improvement? Improve prediction of climate variability & change? Ozone, climate & air quality impacts of atmospheric composition? Change in water cycle dynamics? Predict & mitigate natural hazards from Earth surface change? Carbon cycle & ecosystem change? Climate Variability and Change Atmospheric Composition Carbon Cycle and Ecosystems Weather Water and Energy Cycle Earth Surface and Interior Science Questions and Focus Areas

Motivation and Background: 

Motivation and Background Tornado lead time -12 min national average, improvement stagnant Lightning strikes responsible for >500 injuries per year, 90% of victims suffer permanent disabilities and long term health problems, chiefly neurological in nature Lightning responsible for 80 deaths per year (second leading source after flooding) Aviation weather- airport operations, enroute savings $25M/yr In-cloud lightning lead time of impending ground strikes, often 10 min or more … forest fire initiation, utility crew deployment, NEXRAD coverage gaps, improved precipitation estimates, hail/wind/flood detection, …

Slide5: 

WWRP/Tom Keenan

Warn on Forecast Concept: 

Warn-on-forecast (warnings out to 4 hours based upon observations + short term model forecasts) Warn on Forecast Concept Courtesy Kevin Kelleher, SDR Grand Challenges

Slide7: 

Enhanced Thunder 19 July 2005 2000Z – 2012Z NCEP SPC/Schaefer SPC Experimental Product

Slide8: 

F15 SREF 3-hr COMBINED PROBABILITY OF LIGHTNING - Pr (CPTP) >= 1 x Pr (PCPN) >= .01” Uncalibrated probability of lightning SPC Experimental Product

Mapping storm initiation, growth, decay: 

Mapping storm initiation, growth, decay TRMM provides us a huge database of paired lightning, radar, IR and passive microwave observations (training / validation) Over entire tropics & subtropics (generalization) Total lightning increases as storm intensifies – can increase lead time for warning of severe and tornadic storms

Slide10: 

Flash Rate Coupled to Mass in the Mixed Phase Region Cecil et al., Mon. Wea. Rev. 2005 (from TRMM Observations)

Slide11: 

WRF Thunderstorm Premises and Objectives WRF: Weather Research and Forecast Model CRM: Cloud Resolving Model Additional Forecast Interests CI - convective initiation Ti - First lightning (35 dBZ at -15C, glaciation) Tp - Peak flash rate ~ VIL (Mass) Tf - Final lightning Given: Precipitating ice aloft is correlated with LTG rates Mesoscale CRMs are being used to forecast convection CRMs can represent many ice hydrometeors (crudely) Goals: See if WRF can forecast LTG threat, based on ice flux in layers near -15 C.

WRF Thunderstorm and Lightning Forecasts:Methodology : 

WRF Thunderstorm and Lightning Forecasts: Methodology Use high-resolution (2-4 km) WRF simulations to prognose convection Develop diagnostics from model output fields to serve as a proxy for LTG Create 0-12 h forecasts of LTG threat based on WRF data Compare WRF forecasts with actual reflectivity and LTG and other observations using HSV area assets Subjectively and objectively evaluate WRF capabilities for forecasting LTG

WRF Thunderstorm and Lightning Forecasts:10 December 2004 : 

WRF Thunderstorm and Lightning Forecasts: 10 December 2004

WRF Configuration10 December 2004 Case Study: 

WRF Configuration 10 December 2004 Case Study 2 km horizontal resolution 52 vertical levels Dynamics and physics Eulerian mass core Dudhia SW radiation RRTM LW radiation YSU PBL scheme Noah LSM WSM 6-class microphysics scheme Explicit convection 8h forecast initialized at 12 UTC 10 December 2004 with AWIP212 NCEP EDAS analysis Eta 3-h forecasts used for LBC’s Hot-start with ACARS, METARS, NEXRAD velocity Cloud cover 18h forecast valid at 18 UTC 10 Dec 2004

Slide15: 

WRF vs Eta Surface-based CAPE 18h fcst valid 18 UTC Dec 10

WRF Sounding ~ 800 J/kg CAPE: 

WRF Sounding ~ 800 J/kg CAPE

MIPS Sounding ~ 761 J/kg CAPE: 

MIPS Sounding ~ 761 J/kg CAPE Low level lapse rates and low freezing level efficient for converting CAPE to kinetic energy Surface: T=15C, Td=10C Similar CAPE to MIPS, but for different reasons High-res RAMS storm: Max w = 19 m/s UAH MIPS, Kevin Knupp

Slide18: 

WRF vs Eta 3h Regional Precip. 21h fcst valid 21 UTC 10 Dec 2004

Slide19: 

WRF vs Eta 3h Local Precip. 21h fcst valid 21 UTC 10 Dec 2004 Question: Any lightning, when was it, What was WRF reflectivity at -15 C?

Slide20: 

WRF Reflectivity (dBZ) at -15 C (4.0 km) 1200 UTC forecast valid at 18:50 UTC 10 Dec 2004

Slide21: 

x1: Reflectivity (dBZ), Temperature (°C), and Pressure (hPa) 1200 UTC forecast valid at 18:50 UTC 10 Dec 2004 Max dBZ< 40 dBZ

Slide22: 

x2: Reflectivity (dBZ), Temperature (°C), and Pressure (hPa) 6h 50m forecast valid at 18:50 UTC 10 Dec 2004 Max dBZ~50 dBZ; wmax only 4 m/s; But no hail reaches the surface

Slide23: 

Ground-truth Report of Dime-Size Hail Owens Crossroads, AL, 10 Dec 2004

Slide24: 

At 17:55 IC fl. rate ~ 3/minute in southern cell No IC’s in northern cell at 17:55 No CG’s in either cell for 20 minutes centered on 17:55 Only 3 CG’s detected for duration of storms

High-res RAMS “Validation” Run: 

High-res RAMS “Validation” Run 500 m horizontal resolution Height, Dz is variable, from 250 m at bottom to 750 m at 20 km height Domain 75 km x 75 km x 24.5 km Time, Dt = 4 s, five acoustic steps between Smagorinsky subgrid mixing scheme 5-class precipitating hydrometeors: Rain, snow, aggregates, graupel, hail Initialized with 3K warm bubble, radius=12 km at z=0 120 min simulation, initiation effects dominate until t=60 min Reflectivity Note: wmax reaches 19 m/s

RAMS Graupel Cross-Section: 

RAMS Graupel Cross-Section 500 m horizontal resolution Height, Dz is variable, from 250 m at bottom to 750 m at 20 km height Domain 75 km x 75 km x 24.5 km Time, Dt = 4 s, five acoustic steps between Smagorinsky subgrid mixing scheme 5-class precipitating hydrometeors: Rain, snow, aggregates, graupel, hail Initialized with 3K warm bubble, radius=12 km at z=0 120 min simulation, initiation effects dominate until t=60 min Graupel

RAMS Hail Cross-Section: 

RAMS Hail Cross-Section 500 m horizontal resolution Height, Dz is variable, from 250 m at bottom to 750 m at 20 km height Domain 75 km x 75 km x 24.5 km Time, Dt = 4 s, five acoustic steps between Smagorinsky subgrid mixing scheme 5-class precipitating hydrometeors: Rain, snow, aggregates, graupel, hail Initialized with 3K warm bubble, radius=12 km at z=0 120 min simulation, initiation effects dominate until t=60 min Hail Note: some hail reaches surface

WRF Thunderstorm and Lightning Forecasts:31 May 2004: 

WRF Thunderstorm and Lightning Forecasts: 31 May 2004

WRF Configuration31 May 2004 Case Study: 

WRF Configuration 31 May 2004 Case Study 2 km horizontal resolution 52 vertical levels Dynamics and physics Eulerian mass core Dudhia SW radiation RRTM LW radiation YSU PBL scheme Noah LSM WSM 6-class microphysics scheme Explicit convection 6 h forecast initialized at 00 UTC 31 May 2004 with AWIP212 NCEP EDAS analysis Eta 3-h forecasts used for LBC’s Hot-start with ACARS, METARS, NEXRAD velocity

LMA Source Density: 

LMA Source Density

WRF (vr) Graupel Flux at -15C:31 May 2005: 

WRF (vr) Graupel Flux at -15C: 31 May 2005

WRF (Z) Graupel Flux at -15C:31 May 2005: 

WRF (Z) Graupel Flux at -15C: 31 May 2005

TRMM-based Z-L Transforms Reflectivity Profile to Flash Rate: 

TRMM-based Z-L Transforms Reflectivity Profile to Flash Rate Linear regression of LIS flash rate vs 6 km and 9 km dBZ

1. WRF forecasts of deep convection are useful, but of variable quality - Timing and intensity of convection are depicted fairly well - Location and morphology of storm systems sometimes wrong2. WRF convection is deep enough, with sufficient reflectivity, to suggest lightning3. WRF updraft strengths on 2-4 km grids often too weak, relative to observed weather and high-res RAMS simulations4. WRF microphysics still too simple; need more ice categories5. Finer model mesh may improve updraft representation, and hydrometeor amounts6. Biggest limitation is likely errors in initial mesoscale fields : 

1. WRF forecasts of deep convection are useful, but of variable quality - Timing and intensity of convection are depicted fairly well - Location and morphology of storm systems sometimes wrong 2. WRF convection is deep enough, with sufficient reflectivity, to suggest lightning 3. WRF updraft strengths on 2-4 km grids often too weak, relative to observed weather and high-res RAMS simulations 4. WRF microphysics still too simple; need more ice categories 5. Finer model mesh may improve updraft representation, and hydrometeor amounts 6. Biggest limitation is likely errors in initial mesoscale fields Conclusions

Expand catalog of simulation cases to obtain robust statistics - Collaboration with SPoRT (NASA-NOAA Short-term Prediction Research and Transition Center) - Collaboration with Hazardous Weather Testbed (HWT)2. Develop quantitative metrics for LTG forecasts - Need link between graupel flux and lightning probabilities, rates3. Compare LTG threat parameter against environmental variables- CAPE, LI, etc.4. Enhance accuracy of WRF forecasts5. Hot start WRF applied to archived North Alabama WES cases: 

Expand catalog of simulation cases to obtain robust statistics - Collaboration with SPoRT (NASA-NOAA Short-term Prediction Research and Transition Center) - Collaboration with Hazardous Weather Testbed (HWT) 2. Develop quantitative metrics for LTG forecasts - Need link between graupel flux and lightning probabilities, rates 3. Compare LTG threat parameter against environmental variables- CAPE, LI, etc. 4. Enhance accuracy of WRF forecasts 5. Hot start WRF applied to archived North Alabama WES cases Future Work

Web Sites: 

Web Sites http://weather.msfc.nasa.gov (SPoRT, Workshops) http://branch.nsstc.nasa.gov (North AL, DC Metro LMA) http://thunder.msfc.nasa.gov (LI/OTD)