Glendale
Glendale, Arizona, USA
- Training fall for MetDetect
- Human-verified · Marc Fries
- Meteorites recovered
- Fireball (UTC)
- 2018-07-27 03:22:50 ± 30 s video
- Observed by
- How it was observed
- Radar searched
- KIWA
- Meteorites
- Glendale (L6) Meteoritical Bulletin entry
- MetDetect
- MetDetect RUN-42 training only, not in the headline numbers · radar KIWA
How it was observed
- Fireball camerasOne American Meteor Society network camera, run by Mike Farmer, recorded it; no other dedicated camera found.American Meteor Society blog post, 'Meteorite fall in Arizona on July 27th' (V. Perlerin, Aug 2018)
- WitnessesAMS received only 12 reports, mostly around Phoenix; storms hid it, and one witness heard a sonic boom.American Meteor Society blog post, 'Meteorite fall in Arizona on July 27th'
Fireball cameras are cameras built to record meteors, usually all-sky and calibrated. Casual video is footage or photos from cameras that were not (dash cams, doorbell and security cameras, phones); where a source does not say what kind of camera, it is listed as video of unknown type. Only what a source documents is listed; a missing label means nothing of that kind is documented, not that it did not exist.
Automated detection
MetDetect searched the KIWA radar around the fireball time. This is a training fall: there are fewer than 10 hand-identified radar cells on KIWA, too few for the headline numbers, so the fall only helps train MetDetect for the other falls (never for itself). Its answer still contains 5 of the 9 hand-identified cells, but is mostly other echo.
- 5of the 9 hand-identified cells on KIWA found
- 4hand-identified cells missed
- 23cells reported that were not identified by hand
- 18%of the cells it reports were identified by hand (purity)
MetDetect RUN-42, 12 Nov 2026. Leave one fall out: every learned part was trained without the fall it is scored on. 30 of the 32 falls are the test set (dual-polarization radar, at least 10 hand-identified radar measurements on the radar searched); the other 2 are used for training only and are not in the headline numbers. The method was developed on these falls, so these are development results, not a blind test. This fall is not in the headline numbers. The hand identification is the reference these numbers are graded against.
Dark flight
After the fireball goes dark, meteorites fall freely through the wind. Each radar cell is flown down to the ground through the winds of that day; the coloured areas show where stones of each mass most likely landed, and the black line runs from light to heavy stones (solid where the radar traced it, dashed where extended). The map shows both: (a) from MetDetect's cells on KIWA, (b) from the radar cells identified by hand by Marc Fries on every NEXRAD radar labelled, in the same frame and colour scale. Each panel gives its search area for stones of 10 g and up (the 90% regions) and the recovered stones of 10 g and up inside it. Because MetDetect's first pick was not the debris, panel (a) shows where the stones of that other echo would land, if anywhere. It is not a search area for this fall.
Pixel estimate, computed by P. Shober with MetDetect: winds from ERA5 (ECMWF), masses at the fall's assumed bulk density (on the colour bar). Preliminary: it does not fit the fireball itself, so the heavy end of the line is an extrapolation. Stars mark recovered meteorites; positions from Meteoritical Bulletin 110. They are drawn for comparison only and were never used to make the maps. Counted stones are those of 10 g and up, or of unpublished mass. Villages and rivers: © OpenStreetMap contributors.
Human identification
The radar signatures of this fall were identified by hand by Marc Fries. Read the original report (archived copy).
Meteorites
- Glendale · L6
Publications
No papers on this fall yet that we know of.
Know of a paper or a recovered meteorite we missed? Tell us.