Thunderstorm Tracking and Nowcasting using 3 D Lightning

Thunderstorm Tracking and Nowcasting using 3 D Lightning and Radar Data in Southern Germany Vera Meyer [1] - vera. meyer@zamg. ac. at, H. Höller [2], H. -D. Betz [3] , K. Schmidt [2] [1] Central Institute for Meteorology and Geodynamics, Hohe Warte, Vienna [2] Deutsches Zentrum für Luft und Raumfahrt, Institut für Physik der Atmosphäre, Oberpfaffenhofen, Deutschland [3] Physics Department, University of Munich, Germany 1 Convection Week 2011, Session 3

PROJECT Regio. Ex. AKT www. regioexakt. de Regional Risk of Convective Extreme Weather Events: User-oriented concepts for optimised thunderstorm nowcasting, with focus on the needs of Munich Airport Coordinator: Dr. Nikolai Dotzek MUNICH AIRPORT HEAVY RAIN 15 Juni 2007 40, 5 l/m² zw. 18. 00 – 21. 00 h intense rain hail lightning strikes wind gusts etc. 2 HAIL DAMAGE Boeing 737, Geneva 15 August 2003 MOTIVATION

LINET - Lightning Detection Network - ‘total lightning’ detection - 2008 app. 100 sensors in Central Europe - magnetic field measurements - TOA (time of arrival) method - event-height parameter comprehensive discrimination of ‘cloud’ and ‘cloud-to-ground’ lightning 3 MOTIVATION

LIGHTNING TYPES IC Abbreviations IC (cloud lightning) in-cloud, inter-cloud, intra-cloud, cloud-to-air -15 °C IC IC 0 °C CG (cloud-to-ground) cloud – to – ground TL (total lightning) sum (IC + CG) CG 4 MOTIVATION CG

LINET 3 D-Visualisierung high-precision lightning detection. ‚NORMAL STORM‘ ‚SEVERE STORM‘ cloud lightning cloud-to- ground lightning 5 MOTIVATION

THUNDERSTORM TRACKING and NOWCASTING lightning-cell MUNICH AIRPORT cell track temporal evolution of LIGHTNING cell parameters total lightning cloud-to-ground lightning 14: 45 6 15: 00 15: 15 MOTIVATION 15: 30 time

THUNDERSTORM TRACKING and NOWCASTING lightning-cell MUNICH AIRPORT cell track cell nowcasts temporal evolution of LIGHTNING cell parameters total lightning cloud-to-ground lightning 14: 45 15: 00 15: 15 now prognosis 7 MOTIVATION 15: 30 time

INTRODUCTION GOAL: to assess the usability of 3 D total-lightning data for thunderstorm nowcasting separately and in combination with other data sources (radar) 8 INTRODUCTION

INTRODUCTION GOAL: to assess the usability of 3 D total-lightning data for thunderstorm nowcasting separately and in combination with other data sources (radar) identification tracking prediction 9 INTRODUCTION

INTRODUCTION GOAL: to assess the usability of 3 D total-lightning data for thunderstorm nowcasting separately and in combination with other data sources (radar) identification tracking prediction cell evolution 10 INTRODUCTION

INTRODUCTION GOAL: to assess the usability of 3 D total-lightning data for thunderstorm nowcasting separately and in combination with other data sources (radar) - develop a nowcasting method based on lightning information - develop a method to compare lightning-cell information with information from other data sources (radar) verify lightning-cell properties in case-studies evaluate the statistical information content of 3 D lightning information 11 INTRODUCTION

RESEARCH DOMAIN and OBSERVATION PERIOD May – September 2008 12 INTRODUCTION

NOWCASTING APPROACH ec-TRAM – tracking and monitoring of electrically charged convective cells combines cell informations from independently tracked lightning- and radar-cells 13 METHOD

NOWCASTING APPROACH ec-TRAM – tracking and monitoring of electrically charged convective cells DWD Radar Site Fürholzen (Munich) 2 D reflecitvity maps, low level scan domain [200 km x 200 km] resolution [1 km x 1 km], [5 min] LINET lightning data, nowcast Gmb. H 3 D TOA method in VLF/LF regime, IC/CG discrimination 2 D discharge event maps cell clustering: time interval minimum distance 14 METHOD 3 3 min 6 km

NOWCASTING APPROACH ec-TRAM combines the cell informations of lightning cells and radar cells cell identification parameter (optimized) radar-cells lightning cell: threshold of 1 event lightning data: amplitude |A| > 2. 5 k. A Rad-TRAM [Kober, 2009] radar cell: threshold of 33 d. BZ lightning-cell li-TRAM [Meyer, 2010] cell assignment via spatial overlap 15 METHOD 4

NOWCASTING APPROACH ec-TRAM combines the cell informations of lightning cells and radar cells radar-cells cell identification parameter lightning cell: Rad-TRAM [Kober, 2009] threshold of 1 event radar cell: threshold of 33 d. BZ ec-cells lightning-cell ec-TRAM [Meyer, 2010] li-TRAM [Meyer, 2010] cell assignment via spatial overlap 16 METHOD 4

NOWCASTING APPROACH ec-TRAM ec-cells cell track example: ec-TRAM nowcasting map (detail) with cell contours, tracks and prognoses of an electrically charged ‚ec-cell‘. radar cell: Reflectivity map (blue shaded) cell track (white line), actual cell contour (white polygons), cell prognoses for 10 minutes (dark grey polygons), and 20 minutes (light grey polygons) lightning cell: discharge events clustered for 3 minutes (green crosses) actual cell (red polygon) 17 METHOD 5

TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER 25 June 2008 F P South Germany Austria radar sites x Fürholzen x POLDIRAD MUC M R A 18 18 CASE STUDY Munich Airport Munich Regensburg Augsburg

19 19

TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER Example rad-TRAM: temporal evolution of selected parameters radar-cell: -cell area 20 25 June 2008 [km²] CASE STUDY

TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER Example li-TRAM: temporal evolution of selected parameters 25 June 2008 lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 21 CASE STUDY

TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER radar-cell: -cell area [km²] lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 22 25 June 2008 area [km²], discharge frequency [cnt/cell] Example ec-TRAM: temporal evolution of selected parameters 22 CASE STUDY

Example ec-TRAM: temporal evolution of selected parameters radar-cell: -cell area [km²] lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 23 area [km²], discharge frequency [cnt/cell] TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER onset 23 CASE STUDY

Example ec-TRAM: temporal evolution of selected parameters radar-cell: -cell area [km²] lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 24 area [km²], discharge frequency [cnt/cell] TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER cell splitting 24 CASE STUDY

Example ec-TRAM: temporal evolution of selected parameters radar-cell: -cell area [km²] lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 25 area [km²], discharge frequency [cnt/cell] TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER intensification 25 CASE STUDY

Example ec-TRAM: temporal evolution of selected parameters radar-cell: -cell area [km²] lightning-cell: -cell area [km²] - TL [cnt/cell] - CG [cnt/cell] - IC [cnt/cell] 26 area [km²], discharge frequency [cnt/cell] TEMPORAL EVOLUTION of ec-TRAM CELL PARAMETER decease 26 CASE STUDY

VERIFICATION of LIGHTNING-CELL PROPERTIES in CASE-STUDIES lifetime series of ec-cell parameters were complemented with 3 D polarimetric radar data (POLDIRAD) - not shown lightning-cell parameters were found to -evolve reasonably according to the current state of knowledge -be in very good agreement with other case studies [Klemp 1987, Williams 1989 and 1999, Goodman 1988, Carey 1996, Lopez 1997, Mazur 1998, Altaraz 2003, Motley 2006, . . . ] à reflect the actual storm dynamic (intensification / weakening) à li-TRAM has reasonable, consistent tracking performances (comparable to rad-TRAM) [Meyer, 2010] 27 VERIFICATION

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell area [km²] Lifetime = 40 min 28 2 CASE STUDIES

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell growth 3 2 I cell area [km²] Lifetime = 40 min 29 2 CASE STUDIES

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 mature stage cell area [km²] Lifetime = 40 min 30

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell dissipation 4 5 6 E cell area [km²] Lifetime = 40 min 31 2 CASE STUDIES

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell area [km²] Lifetime = 40 min 32 2 CASE STUDIES cell area [km²] Lifetime = 145 min

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell area [km²] cell growth 4 I Lifetime = 40 min 33 2 CASE STUDIES 5 23 cell area [km²] Lifetime = 145 min

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell area [km²] Lifetime = 40 min 34 2 CASE STUDIES mature stage cell area [km²] Lifetime = 145 min

PARAMETER CORRELATIONS of 2 LIGHTNING-CELL TRACKS 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 cell area [km²] cell dissipation -4 -2 E Lifetime = 40 min 35 -5 2 CASE STUDIES -6 -3 cell area [km²] Lifetime = 145 min

lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS TL mean TL MEAN IC mean IC MEAN TL fit 1 + 2 PARAMETER MEANS lightning frequency versus cell area -10 km² area intervals -10 200 completely assessed lightning-cell entries cell area [km²] 36 LIGHTNING-STATISTICS

lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS TL mean TL MEAN IC mean IC MEAN TL fit 1 + 2 160 km² cell area [km²] 37 LIGHTNING-STATISTICS

TL MEAN IC MEAN TL fit 1 + 2 160 km² cell area [km²] 38 IC mean discharge height per cell [km] lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS IC MEAN height fit 1 + 2 LIGHTNING-STATISTICS cell area [km²]

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS TL MEAN IC MEAN TL fit 1 + 2 160 km² cell area [km²] 39 IC mean discharge height per cell [km] lightning frequency per cell [cnt/cell] NO ARTIFACT of the ALGRITHM IC mean height IC MEAN height fit 1 + 2 LIGHTNING-STATISTICS 160 km² cell area [km²]

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS TL mean MEAN IC mean MEAN TL fit 1 + 2 160 km² cell area [km²] 40 IC mean discharge height per cell [km] lightning frequency per cell [cnt/cell] NO INFORMATION about TEMPORAL CELL EVOLUTION IC mean height IC MEANheight fit 1 + 2 LIGHTNING-STATISTICS 160 km² cell area [km²] 9

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS I E cell area [km²] Lifetime = 40 min 41 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 I E cell area [km²] Lifetime = 145 min LIGHTNING-STATISTICS 10

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS I E cell area [km²] Lifetime = 40 min 42 25 May 2008 lightning-cell No 6 lightning frequency per cell [cnt/cell] 30 May 2008 lightning-cell No 137 I E cell area [km²] Lifetime = 145 min LIGHTNING-STATISTICS 10

frequency [-] FREQUENCY DISTRIBUTION of LIFE-TIMES short-lived cells long-lived cells lightning-cell lifetime [min] 43 LIGHTNING-STATISTICS

LIFETIME REGIMES short-lived [ 15 min – 75 min ] ‚SINGLE CELLS‘ - lowly organized - simply structured: - simple life-cycles: long-lived 1 updraft + 1 downdraft growth – short maturity – decease [ ≥ 80 min ] ‚MULITCELLS‘, ‚SUPERCELLS‘ - highly organized - complexly structured - complex life-cycles: growth – elongated (fluctuating) maturity – decease 44 LIGHTNING-STATISTICS

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS total lightning relative frequency lightning frequency per cell [cnt/cell] RELATIVE AMOUNT to STATISTICAL MEAN long-lived cells shortlived cell area [km²] 45 LIGHTNING-STATISTICS cell area [km²]

lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS total lightning cell type short-lived cells shortlived h wt o gr e as e ec d cell area [km²] 46 DISCUSSION

lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS total lightning long-lived cells cell type y urit t ma h wt o long-lived cells gr e as e ec d cell area [km²] 47 DISCUSSION

TL lightning totalmean IC mean TL fit 1 + 2 SCATTER! cell area [km²] 48 48 IC mean discharge height per cell [km] Lightning frequency per cell [1/km²] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS DISCUSSION IC mean height fit 1 + 2 SCATTER! cell area [km²]

CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS lightning frequency per cell [cnt/cell] INFORMATION about STORM TYPE (lifetime, intensity) and TEMPORAL EVOLUTION! long-lived cell type short-lived shortlived cell area [km²] 49 DISCUSSION long-lived cells

THUNDERSTORM TRACKING and NOWCASTING lightning-cell MUNICH AIRPORT cell track cell nowcasts temporal evolution of LIGHTNING cell parameters total lightning cloud-to-ground lightning 14: 45 15: 00 15: 15 now prognosis 50 Ad MOTIVATION 15: 30 time

TL MEAN IC MEAN TL fit 1 + 2 cell area [km²] 51 IC mean discharge height per cell [km] lightning frequency per cell [cnt/cell] CORRELATION STATISTICS of LIGHTNING-CELL PARAMETERS DISCUSSION IC MEAN height fit 1 + 2 3 D information from 2 D cell tracking! cell area [km²]

CONCLUSION GOAL: to assess the usability of 3 D total-lightning data for thunderstorm nowcasting separately and in combination with other data sources (radar) 3 D lightning information with in-cloud and cloud-to-ground lightning discrimination provides useful information about the storm dynamic and developement and have the capacity to nowcast cell trends from the cell history 52 CONCLUSION and OUTLOOK 11

OUTLOOK - test the usability of (specified) normalized cell life-cycles to derive trend prognoses - use cell trends from cell history to add trend prognoses to local prognoses - test the quality of trend prognoses - investigate cell parameter correlations with other data sources (3 D radar, satellite, . . . ) 53 CONCLUSION and OUTLOOK 11

OTHER possible APPLICATIONS - parameterization of TL frequency with IC/CG ratio and mean IC height for modelling [Price and Rind 1992, Allen and Pickering 2002] - simulation of thunderstorm life-cycles with realistic discharge characteristics ? 160 km²? 54 CONCLUSION and OUTLOOK

IC mean height fit 1 + 2 THANK YOU 160 km² cell area [km²] 55 IC mean discharge height per cell [km] lightning frequency per cell [cnt/cell] TL mean IC mean TL fit 1 + 2 160 km² vera. meyer@zamg. ac. at cell area [km²]

Literature K. Kober and A. Tafferner. Tracking and nowcasting of convective cells unsing remote sensing data from radar and satellite. Meteorologiesche Zeitschrift, 10(1): 75 -84, 2009 V. Meyer, H. Höller, K. Schmidt, and H. -D. Betz. Temporal evolution of total lightning and radar parameters of thunderstorms in southern Germany and its benefit for nowcasting. Proceedings: 5 th European Conference on Severe Storms, 2009 V. Meyer (2010): Thunderstorm Tracking and Monitorin on the Basis of Three Dimenional Lightning Data and Conventional and Polarimetric Radar Data. Dissertation, LMU München: Faculty of Physics http: //edoc. ub. uni-muenchen. de/12102/ 56 vera. meyer@zamg. ac. at

POLARIMTRIC INFORMATION POLDIRAD RHI 20080625, hydrometeorclassifications, 16: 42 h, azimuth = 52 ° sounding munich: 0 ° at 3. 5 km, Tropopause at 10 km 57 57

ZEITLICHE ENTWICKLUNG VON ec-ZELL PARAMETERN example: radar-cell: -cell area [km²] lightning-cell: -cell area - TL - CG - IC [km²] [cnt/cell] polar. radar data - hydrometeors d. BZ cell graupel/hail light rain H = 4 km ground heavy rain 58 area [km²], discharge frequency [cnt/cell] 25 June 2008