storm tracking & typing for lightning observations kristin calhoun, don macgorman, ben herzog
TRANSCRIPT
![Page 1: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/1.jpg)
Storm tracking & typing for lightning observations
Kristin Calhoun, Don MacGorman, Ben Herzog
![Page 2: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/2.jpg)
Storm tracking: Segmotion / k-means• Identifies segments (or storm cells/clusters) of image &
estimates motion
• Extracts properties of each cell • Size of cell (# of pixels)• Pull data from all grids, including radar, lightning, and derived
properties such as maximum expected size of hail (MESH)
• Moves grid based on motion field
• Tracking: Lightning vs. Reflectivity vs. Hybrid
• Want a relatively stable area, without drops or mismatches
• Past track given by centroid of area, if loses section or adds will change centroid location and can account for difference in track
![Page 3: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/3.jpg)
original: reflectivity at -10 C
![Page 4: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/4.jpg)
thresholds & median filter applied
![Page 5: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/5.jpg)
scale 0 (200 km2)
![Page 6: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/6.jpg)
scale 1 (600 km2)
![Page 7: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/7.jpg)
scale 2 (1000 km2)
![Page 8: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/8.jpg)
Flash Extent Density vs Flash Rate vs Flashes per unit area?
• For assimilation: • # of flash initiations (single point: lat / lon) • # of flashes traversing a grid cell, • or some measure of the summed density/duration of
flashes at a particular grid cell?
![Page 9: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/9.jpg)
Flash Extent Density vs Flash Ratevs Flashes per unit area?
![Page 10: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/10.jpg)
Flash Extent Density vs Flash Rate
•GLM scale
![Page 11: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/11.jpg)
Flash Extent Density vs Flash Rate vs Flashes per unit area?
• For assimilation: • # of flash initiations, • # of flashes traversing a grid cell, • or some measure of the summed density/duration of flashes at a
particular grid cell?
• For observations / lightning jump: • # of flashes per cell, • the max density of flashes in a cell, • or the # of flashes per unit area of a cell
• Are the answers to these questions different for different types of storms or at different resolutions/spatial scales?
• How does this translate to pGLM and ultimately GLM?
![Page 12: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/12.jpg)
6 years of data (2007-2012) 3 domains (OKLMA, NALMA, DCLMA)
![Page 13: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/13.jpg)
Speed tests! active weather week over 3 LMA domainsOn NSSL machine idealized for real-time processing:• 21 single radars, processing: ~31 hrs/each• 3 LMA domains, processing: ~29 hrs/each• Multi-radar/sensor merger & storm tracking:
~155 hrs
Onto supercomputing:• ~67% faster + more parallel computing
opportunities!• Run multiple weeks at the same time• Optimizing WDSS2 for supercomputing
Can run in parallel
![Page 14: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/14.jpg)
Storm typing• Decision tree: based on training set from number of
years across the CONUS (Hobson et al., 2012)
http://dx.doi.org/10.1016/j.atmosres.2012.03.004
![Page 15: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/15.jpg)
short-lived, multi-cell, supercell
Storm typing: 19 May 2010, NW OK
![Page 16: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/16.jpg)
a bit later: 2 supercell storms
Storm typing: 19 May 2010, NW OK
![Page 17: Storm tracking & typing for lightning observations Kristin Calhoun, Don MacGorman, Ben Herzog](https://reader036.vdocuments.mx/reader036/viewer/2022062515/56649f3f5503460f94c5fc54/html5/thumbnails/17.jpg)
Applying results to Data Assimilation:• 3DVAR framework – real-time in the Hazardous Weather Testbed
every 5 min (w/Jidong Gao)
• Parallel work by Ted Mansell, Blake Allen, and Alex Fierro utilizing EnKF and pGLM flash density
Radar #2
Radar #2
Radar #3
Radar #3
Radar #1
Radar #1
Total Lightning
satellite
Surface obs