3421 Rate this article:

If a Meteor Vaporizes in the Atmosphere, Does It Make a Sound?

Jim Pendleton Tuesday, November 8, 2022

Jim Pendleton is a Professional Services Engineer at NV5 Geospatial

Himawari blog image

In this screen capture from Himawari 8 Band 3 data (0.64 microns), the vapor trail of the large meteor, or bolide, (after it presumably met its demise about 2 minutes before the image was taken), is highlighted in the diamond and magnified inset image. The linear feature indicated by the arrow is either a shadow cast by the sun from the vapor trail onto clouds lower in the atmosphere, or a different part of the debris trail left higher in the atmosphere. Watch the video, below, and decide for yourself.

As a follow-up to another blog on the topic of the long-lasting atmospheric pressure wave generated by the eruption of the Hunga Tonga-Hunga Ha’apai volcano, I thought it would be interesting to learn if objects entering the atmosphere from the other direction can show a similar signal in weather satellite data.

I recently received a link to a CBC News article from 2019 by Nicole Mortillaro. It describes a large meteor, or bolide, impacting the Earth’s atmosphere over the Bering Sea on December 18, 2018. A second article from the UK’s DailyMail includes a video from the Japan Meteorological Agency’s Himawari 8 weather satellite of this event. Both articles note that scientists estimate the bolide exploded with the energy of nearly 200 kilotons. This is a far cry from the Hunga Tonga volcano’s estimated energy release of 10 megatons, and the mechanism – and direction – of atmospheric disturbance is very different, of course. However, according to these articles, scientists estimate only a handful of bolides of this size impact Earth’s atmosphere, per century, and archival data is available from various satellite sources for this date, so I considered the investigation to be worth a shot.

NASA’s Jet Propulsion Center’s Center for Near Earth Object Studies “Fireballs” web page provided the coordinates and date of this event. The estimated date and time are December 18, 2018 at 23:48:20 UTC at the location 56.9 degrees North and 172.4 degrees East, between Alaska and Russia over the Bering Sea.

This is at least 200 miles from any population center. It was daylight in that location, the sky was overcast for hundreds of miles in any direction, and the bolide likely vaporized 70 to 100 km high in the thermosphere. Therefore, it is unlikely anyone on Earth’s surface witnessed phenomena associated with the event first-hand.

Having no luck finding any features with the freely available GOES-16 and GOES-17 data archives for this date, I requested non-commercial access to the Himawari archival data. This process requires two steps. First, a user must make a request to the Data Integration & Analysis System (DIAS). DIAS was developed and is operated by a project supported by the Ministry of Education, Culture, Sports, Science and Technology, Japan and the University of Tokyo.

After initial registration and acceptance of their terms of use, a user must make a second request specifically for access to the archival Himawari 8 and 9 datasets. They are not exposed directly through the DIAS Data Search and Discovery System’s main web page. Once access was granted and a link was provided, I was able to download a selection of archival data around this event.

First, I downloaded raw, “original data” from Band 9, centered at 6.9 microns, approximately matching the wavelength used in my earlier time-difference imaging to display the shockwave from the Hunga Tonga volcano eruption. Because my initial hope was to identify a spatially- and temporally-extended phenomenon, I requested the full disk imagery, all 10 “segments” per acquisition “timeline”. It was only then that I realized neither IDL® nor ENVI® have the built-in capabilities of reading the raw Himawari Standard Data format.

Fortunately, the Japan Meteorological Society provides a User’s Guide document describing the format. The documentation was straightforward and the IDL code I wrote to read this format is provided below.

		
;  Reference: 
;  Himawari-8/9
;  Himawari Standard Data
;  User's Guide
;  Version 1.3 
;  3 July, 2017 
;  Japan Meteorological Agency 
;  1-3-4 Otemachi, Chiyoda-ku, Tokyo, 100-8122 Japan
;
;Found here: 
;  https://www.data.jma.go.jp/mscweb/en/himawari89/space_segment/hsd_sample/HS_D_users_guide_en_v13.pdf
;
; We use fixed-length block definitions to read the data from
; file, but then convert them to hash storage after input. Among other things,
; fixed-length byte strings are converted to IDL strings.

Pro HimawariStandardBlock1Data__Define
; #1 Basic information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock1Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    TotalHeaderBlocks : 0U, $
    ByteOrder : 0B, $
    SatelliteName_B : BytArr(16), $
    ProcessingCenter_B : BytArr(16), $
    ObservationArea_B : BytArr(4), $
    OtherObservationInfo_B : BytArr(2), $
    ObservationTimeline : 0U, $
    ObservationStartTime : 0D, $
    ObservationEndTime : 0D, $
    FileCreationTime : 0D, $
    TotalHeaderLength : 0UL, $
    TotalDataLength : 0UL, $
    QualityFlag1 : 0B, $
    QualityFlag2 : 0B, $
    QualityFlag3 : 0B, $
    QualityFlag4 : 0B, $
    FileFormatVersion_B : BytArr(32), $
    FileName_B : BytArr(128), $
    Spare : BytArr(40)}
End

Pro HimawariStandardBlock2Data__Define
; #2 Data information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock2Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    NumberOfBitsPerPixel : 0U, $
    NumberOfColumns : 0U, $
    NumberOfLines : 0U, $
    CompressFlagForDataBlock12 : 0B, $
    Spare : BytArr(40) $
  }
End

Pro HimawariStandardBlock3Data__Define
; #3 Projection information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock3Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    Sub_Lon : 0D, $
    CFAC : 0UL, $
    LFAC : 0UL, $
    COFF : 0., $
    LOFF : 0., $
    DistanceFromEarthCenterToVirtualSatellite : 0D, $
    EarthsEquatorialRadius : 0D, $
    EarthsPolarRadius : 0D, $
    F1 : 0D, $
    F2 : 0D, $
    F3 : 0D, $
    Coefficient : 0D, $
    ResamplingTypes : 0U, $
    ResamplingSize : 0U, $
    Spare : BytArr(40) $
  }
End

Pro HimawariStandardBlock4Data__Define
; #4 Navigation information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock4Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    NavigationInformationTime : 0D, $
    SSPLongitude : 0D, $
    SSPLatitude : 0D, $
    DistanceFromEarthsCenterToSatellite : 0D, $
    NadirLongitude : 0D, $
    NadirLatitude : 0D, $
    SunsPosition : DblArr(3), $
    MoonsPosition : DblArr(3), $
    Spare : BytArr(40) $
  }
End

Pro HimawariStandardBlock5Data__Define
; #5 Calibration information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock5Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    BandNumber : 0U, $
    CentralWavelength : 0D, $
    ValidNumberOfBitsPerPixel : 0U, $
    CountValueofErrorPixels : 0U, $
    CountValueOfPixelsOutsideScanArea : 0U, $
    SlopeForCountRadianceConversion : 0D, $
    InterceptForCountRadianceConversion : 0D, $
    CorrectionCoefficientRadianceToBrightnessTemperaturec0 : 0D, $
    CorrectionCoefficientRadianceToBrightnessTemperaturec1 : 0D, $
    CorrectionCoefficientRadianceToBrightnessTemperaturec2 : 0D, $
    CorrectionCoefficientForSensorBrightnessTemperatureToRadiancec0 : 0D, $
    CorrectionCoefficientForSensorBrightnessTemperatureToRadiancec1 : 0D, $
    CorrectionCoefficientForSensorBrightnessTemperatureToRadiancec2 : 0D, $
    SpeedOfLight : 0D, $
    PlanckConstant : 0D, $
    BoltzmannConstant : 0D, $
    Spare : BytArr(40) $
    }
End

Pro HimawariStandardBlock6Data__Define
; #6 Inter-calibration information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock6Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    GSICSCalibrationCoefficientIntercept : 0D, $
    GSICSCalibrationCoefficientSlope : 0D, $
    GSICSCalibrationCoefficienQuadtraticTerm : 0D, $
    RadianceBiasForStandardScene : 0D, $
    UncertaintyOfRadianceBiasForStandardScene : 0D, $
    RadianceForStandardScene : 0D, $
    StartTimeOfGCSICSCorrectionValidityPeriod : 0D, $
    EndTimeOfGSCISCorrectionValidityPeriod: 0D, $
    RaidanceValidityRangeOfGSICSCalibrationCoefficientsUpperLimit : 0., $
    RaidanceValidityRangeOfGSICSCalibrationCoefficientsLowerLimit : 0., $
    FileNameOfGSICSCorrection_B : Bytarr(128), $
    Spare : BytArr(56) $
    }
End

Pro HimawariStandardBlock7Data__Define
; #7 Segment information block
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock7Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    TotalNumberOfSegments : 0B, $
    SegmentSequenceNumber : 0B, $
    FirstLineNumberOfImageSegment : 0U, $
    Spare : BytArr(40) $
    }
End

Pro HimawariStandardBlock8Data__Define
; #8 Navigation correction information block 
; Block 8 has an indeterminate number of extra records for
; shift data, along with a Spare vector that are added
; later.
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock8Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    CenterColumnOfRotation : 0., $
    CenterLineOfRotation : 0., $
    AmountOfRotationalCorrection : 0D, $
    NumberofCorrectionInformationDataForColumnAndLineDirection : 0U $
    }
End

Pro HimawariStandardBlock8ShiftData__Define
; Block 8 may have 0 to N of these shift data
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock8ShiftData, $
    LineNumberAfterRotation : 0U, $
    ShiftAmountForColumnDirection : 0., $
    ShiftAmountForLineDirection : 0. $
    }
End

Pro HimawariStandardBlock9Data__Define
; #9 Observation time information block 
; Block 9 has an indeterminate number of extra records for
; observation time data, along with a Spare vector that are added
; later.
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock9Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    NumberOfObservationTimes : 0U $
    }
End

Pro HimawariStandardBlock9ObservationTimeData__Define
; Block 9 may have 0 to N of these observation time data
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock9ObservationTimeData, $
    LineNumber : 0U, $
    ObservationTime : 0D $
    }
End

Pro HimawariStandardBlock10Data__Define
; #10 Error information block 
; Block 10 has an indeterminate number of extra records for
; error information data, along with a Spare vector that are added
; later.
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock10Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0L, $
    NumberOfErrorInformationData : 0U $
    }
End

Pro HimawariStandardBlock10ErrorInformationData__Define
; Block 10 may have 0 to N of these error information data
Compile_Opt StrictArr, Hidden
!null = {HimawariStandardBlock10ErrorInformationData, $
    LineNumber : 0U, $
    NumberOfErrorPixelsPerLine : 0U $
    }
End

Pro HimawariStandardBlock11Data__Define
; #11 Spare block 
Compile_Opt Strictarr, Hidden
!null = {HimawariStandardBlock11Data, $
    HeaderBlockNumber : 0B, $
    BlockLength : 0U, $
    Spare : BytArr(256)  $
    }
End

Pro HimawariStandardBlock__Define
Compile_Opt StrictArr
On_Error, 2
!null = {HimawariStandardBlock, $
    Inherits OrderedHash, $
    BlockNumber : 0U $
  }
End

Function HimawariStandardBlock::Init, BlockNumber
Compile_Opt StrictArr
On_Error, 2
If (~self.OrderedHash::Init()) then Return, 0
self.BlockNumber = BlockNumber
Return, 1
End

Function HimawariStandardBlock::Read, LUN, Error = Error
Compile_Opt StrictArr
On_Error, 2
Error = !null
Catch, ErrorNumber
If (ErrorNumber ne 0) then Begin
    Error = !error_state.msg + ', Block ' + self.BlockNumber.ToString()
    Return, !false
EndIf
Case self.BlockNumber of
  1 : Data = {HimawariStandardBlock1Data}
  2 : Data = {HimawariStandardBlock2Data}
  3 : Data = {HimawariStandardBlock3Data}
  4 : Data = {HimawariStandardBlock4Data}
  5 : Data = {HimawariStandardBlock5Data}
  6 : Data = {HimawariStandardBlock6Data}
  7 : Data = {HimawariStandardBlock7Data}
  8 : Data = {HimawariStandardBlock8Data}
  9 : Data = {HimawariStandardBlock9Data}
  10 : Data = {HimawariStandardBlock10Data}
  11 : Data = {HimawariStandardBlock11Data}
  Else : Message, 'Invalid block number ' + self.BlockNumber.ToString(), /Traceback
EndCase
; We're using structures to read the data, then converting the info
; into a hash with key/value pairs, instead.  The lowercase version
; of the tag name in the structure becomes the corresponding key,
; with the slight exception of strings.
ReadU, LUN, Data
Tags = Tag_Names(Data)
ForEach Tag, Tags, Index Do Begin
  If (Tag.EndsWith('_B')) then Begin
    ; Structure tags that are defined as fixed-length byte arrays
    ; are converted to IDL strings, and we strip the "_B" from
    ; the end of the tag name for use as the key.
    self[(Tag.Split('_'))[0].ToLower()] = String(Data.(Index))
  EndIf Else Begin
    self[Tag.ToLower()] = Data.(Index)
  EndElse
EndForEach
; Blocks 8 through 10 have an indeterminate number of additional records, 
; followed by an allocation of spare data.
Case self.BlockNumber of
  8 : Begin
    ShiftData = List()
    Data = {HimawariStandardBlock8ShiftData} 
    Tags = Tag_Names(Data)
    For I = 0L, Long(self['numberofcorrectioninformationdataforcolumnandlinedirection']) - 1 Do Begin
      ReadU, LUN, Data
      Items = OrderedHash()
      ForEach Tag, Tags, Index Do Begin
        Items[Tag.ToLower()] = Data.(Index)
      EndForEach
      ShiftData.Add, Items
    EndFor
    self['shiftdata'] = ShiftData
    Spare = BytArr(40)
    ReadU, LUN, Spare
    self['spare'] = Spare
    End
  9 : Begin
    ObservationTimes = List()
    Data = {HimawariStandardBlock9ObservationTimeData}
    Tags = Tag_Names(Data)
    For I = 0L, Long(self['numberofobservationtimes']) - 1 Do Begin
      ReadU, LUN, Data
      Items = OrderedHash()
      ForEach Tag, Tags, Index Do Begin
        Items[Tag.ToLower()] = Data.(Index)
      EndForEach
      ObservationTimes.Add, Items
    EndFor
    self['observationtimes'] = ObservationTimes
    Spare = BytArr(40)
    ReadU, LUN, Spare
    self['spare'] = Spare
    End
  10 : Begin
    ErrorInformation = List()
    Data = {HimawariStandardBlock10ErrorInformationData}
    Tags = Tag_Names(Data)
    For I = 0L, Long(self['numberoferrorinformationdata']) - 1 Do Begin
      ReadU, LUN, Data
      Items = OrderedHash()
      ForEach Tag, Tags, Index Do Begin
        Items[Tag.ToLower()] = Data.(Index)
      EndForEach
      ErrorInformation.Add, Data
    EndFor
    self['errorinformation'] = ErrorInformation
    Spare = BytArr(40)
    ReadU, LUN, Spare
    self['spare'] = Spare
    End
  Else :
EndCase
Return, !true
End

Pro HimawariStandard::Cleanup
Compile_Opt StrictArr
On_Error, 2
If (self.LUN ne 0) then Begin
    ; Close the data file
    Free_LUN, self.LUN, /Force
    self.LUN = 0
EndIf
End

Pro HimawariStandard::GetProperty, $
  Image_Data = Image_Data, $
  Metadata = Metadata
Compile_Opt StrictArr
On_Error, 2
If (Arg_Present(Image_Data)) then Begin
  ; Return a copy of the image data.
  Image_Data = *self.pImage
EndIf
If (Arg_Present(Metadata)) then Begin
  ; Return the complete hash of blocks 1 through 11
  Metadata = self.Blocks
EndIf
End

Function HimawariStandard::Open, File, Error = Error
Compile_Opt StrictArr
On_Error, 2
Error = !null
If (File ne !null) then Begin
  self.File = File
  Status = self.OpenFile(Error = Error)
  Return, Status
EndIf
Return, !False
End

Function HimawariStandard::Init, $
  File = File, Error = Error
Compile_Opt StrictArr
On_Error, 2
Error = !null
If (File ne !null) then Begin
  Status = self.Open(File, Error = Error)
  Return, Status
EndIf
Return, 1
End

Function HimawariStandard::OpenFile, Error = Error
Compile_Opt StrictArr
On_Error, 2
Error = !null
Catch, ErrorNumber
If (ErrorNumber ne 0) then Begin
  Catch, /Cancel
  Error = !error_state.msg
  If (self.LUN ne 0) then Begin
    Free_LUN, self.LUN, /Force
  EndIf
  self.LUN = 0
  Return, !false
EndIf
If (~File_Test(self.File)) then Begin
  Error = 'File does not exist'
  Return, !false
EndIf
; Check if byte order of the data matches the OS and use SWAP_ENDIAN on OpenR, if necessary.
OpenR, LUN, self.File, /Get_LUN
DummyBuffer = BytArr(5) ; Skip the first 5 bytes
ReadU, LUN, DummyBuffer
ByteOrder = 0B
ReadU, LUN, ByteOrder
If (ByteOrder eq 1 && (!d.name).ToUpper() eq 'WIN') || $
   (ByteOrder eq 0 && (!d.name).ToUpper() eq 'UNIX') then Begin
  Free_LUN, LUN, /Force
  OpenR, LUN, self.File, /Get_LUN, Swap_Endian = 1
EndIf Else Begin
  ; No byte-swapping, rewind the file pointer.
  Point_LUN, LUN, 0
EndElse
self.LUN = LUN
Return, !true
End

Function HimawariStandard::Read, Error = Error
Compile_Opt StrictArr
On_Error, 2
self.Blocks = OrderedHash()
For I = 1, 11 Do Begin
  Block = Obj_New('HimawariStandardBlock', I)
  Status = Block.Read(self.LUN, Error = Error)
  If (~Status) then Return, !false
  ; Notice that we're using a hash rather than a list to store the blocks
  ; so we "index" using 1-based integer key values 1 through 11 corresponding to
  ; the defined block numbers instead of list index values 0 through 10.
  self.Blocks[I] = Block
EndFor
; "Block 12" has no metadata, only image data, so we simply
; put the image into a pointer, instead of constructing more hash info.
Image = UIntArr(self.Blocks[2, 'numberofcolumns'], self.Blocks[2, 'numberoflines'])
; Technically, we should check the compression flag from Block 2 rather than 
; assuming the image in uncompressed.  What is the compression scheme if it
; is compressed?
ReadU, self.LUN, Image
self.pImage = Ptr_New(Image)
Free_LUN, self.LUN, /Force
self.LUN = 0
Return, !true
End


Function HimawariStandard::RadianceToBrightnessTemperature, RadianceData, Metadata, $
  Replace_Bad = ReplaceBad
; Raidance in W/(m^2 steradian micron) to brightness temperature
; Note: This is a static method
Compile_Opt StrictArr, Static
On_Error, 2
Exceptsave = !except
; Radiance data expected to come from ::CountToRadiance where "bad" pixels
; have been replaced set to 0.
Bad = Where(RadianceData eq 0, NBad, Complement = Good)
M = Metadata[5]
!except = 0
Num = (M['planckconstant'] * M['speedoflight'])/(M['boltzmannconstant']*M['centralwavelength'])
Denom = ALog2((2.*M['planckconstant']*M['speedoflight']^2)/((M['centralwavelength']^5)*RadianceData) + 1.d)
Te = Num/Denom
Tb = M['correctioncoefficientradiancetobrightnesstemperaturec0'] + $
     M['correctioncoefficientradiancetobrightnesstemperaturec1']*Te + $
     M['correctioncoefficientradiancetobrightnesstemperaturec2']*Te^2
If (NBad ne 0) then Tb[Bad] = ReplaceBad ne !null ? ReplaceBad : Min(Tb[Good], /NAN)
!null = Check_Math()
!except = ExceptSave
Return, Tb
End


Function HimawariStandard::CountToRadiance, ImageData, Metadata
; counts to W/(m^2 steradian micron)
; Note: this is a static method
Compile_Opt StrictArr, Static
On_Error, 2
Bad = Where(ImageData ge 65534, NBad)
M = Metadata[5]
RadianceData = (ImageData * M['slopeforcountradianceconversion'] + $
  M['interceptforcountradianceconversion']) > 0
If (NBad ne 0) then RadianceData[Bad] = 0
Return, RadianceData
End


Pro HimawariStandard__Define
Compile_Opt StrictArr
On_Error, 2
!null = {HimawariStandard, $
   Inherits IDL_Object, $
   File : '', $
   LUN : 0L, $
   Blocks : Obj_New(), $
   pImage : Ptr_New() $
   }
End

Using this object class is straightforward, for example:


h = himawaristandard(file='path-to-.dat-file')
status = h.read(error = error)
if (~status) then … ; this is an error state
imagedata = h.image_data
metadata = h.metadata

Two static methods, HimawariStandard::CountToRadiance and HimawariStandard::RadianceToBrightnessTemperature, are provided to convert from the raw count values. Please contact the author with any suggested corrections to these methods.

The image metadata is stored in an IDL ordered hash that is keyed on block numbers 1 through 11, as described in the User’s Guide.

“Block 12”, the raw, unsigned integer image pixel data, is returned through the Image_Data property of the object, instead. The raw data is inverted in Y, in an IDL sense, and references to pixel locations in the metadata begin at pixel (1,1) in the Himawari Standard coordinate system.

Be aware that the resolution of each wavelength band is defined separately as 0.5, 1, or 2 km per pixel at sub-satellite point (SSP), meaning care must be taken when resampling to produce false-color imagery from multiple bands.

The downloads are provided by DIAS in a TAR format with the individual images stored in a “bzip2” format, under the directory tree organized by timeline. On Windows, I used the 7-Zip application to open the archive and extract the “.bz2” files to disk, then Cygwin’s “bunzip2” executable to expand each “.bz2” file into its final Himawari Standard format.

I wrote an additional utility which converts the count data to radiance for each image segment, then combines the segments into a single image array, oriented the “IDL way”. The combined full-disk radiance-valued image, along with an IDL List object containing the original metadata from the individual segments, is written to an IDL SAVE-format file in the same directory as the .DAT files, named using the original components of the root file names.


pro himawari_construct_full_disk, dir_local
compile_opt strictarr
on_error, 2
dir = dir_local ne !null ? dir_local : '\Bering Sea Meteor\Himawari-8\All Bands\bz2s'
files = (file_search(filepath(root = dir, 'HS_*.DAT'))).sort()
bases = file_basename(files)
baselist = list()
foreach base, bases do baselist.add, (base.split('_'))[0:6].join('_')
disks = (baselist.toarray()).uniq()
foreach disk, disks do begin
  outfile = filepath(disk + '.sav', root = dir)
  if (file_test(outfile)) then continue
  imlist = list()
  metadatalist = list()
  ysize = 0L
  files = (file_search(filepath(root = dir, disk + '*.DAT'))).sort()
  foreach file, files, index do begin
    h = himawaristandard(file=file)
    status = h.read()
    if (~status) then continue
    imagedata = h.image_data
    bad = where(imagedata ge 65534, nbad)
    metadata = h.metadata
    metadatalist.add, metadata
    radiancedata = HimawariStandard.CountToRadiance(imagedata, metadata)
    if (nbad ne 0) then radiancedata[bad] = 0
    imlist.add, reverse(radiancedata, 2)
    ysize += (radiancedata.dim)[1]
  endforeach
  radiance = dblarr(((imlist[0]).dim)[0], ysize)
  for i = 0, imlist.count() - 1 do begin
    ; Technically, this should allow for any missing segments by using the yoffset
    ; info from each image's metadata, but we're working with archival data
    ; so all segments should be available in all bands on all days.
    radiance[0, ysize - ((imlist[0]).dim)[1]*(i + 1)] = imlist[i]
  endfor
  save, radiance, metadatalist, /compress, file = outfile
endforeach
end

First, let’s look at the full disk imagery.

Band 3 is 0.5 km resolution (resulting in a full-size image that is 22,000 pixels per side!), so this screen grab is significantly reduced in size. Our area of interest is in the far upper right of the full disk. Continental outlines were generated using the appropriate metadata tags, and IDL’s MAP_SET procedure parameters for a satellite projection.

 map_set, /cont, /satellite, md4['ssplatitude'], md4['ssplongitude'], $
        sat_p = [md4['distancefromearthscentertosatellite']/ $
        md3['earthsequatorialradius'],0,0], $
        /noborder, /noerase, mlinethick = 8, /isotropic, $
        /horizon, color = 'ff0000'x, /hi, $
        xmargin = [1, 8], ymargin = [10,10] ; dependent on scale

This projection is OK when zoomed out, but fails a bit when zoomed in. The margin keyword values are inexact (and are in units of “character size”!). These margins are needed because the tangent to the Earth’s surface does not touch the four sides of the buffer containing the image data. You may wish to tweak these further to achieve a more pleasing and accurate result.

Next, let’s zoom into the area of interest.

Having gone through the exercise of downloading data, writing a file parser, creating the full-disk images, and tweaking the map overlay, did I locate any obvious “shock wave” signal, analogous to the Hunga Tonga volcano’s? No, I did not.

In the following video, the radiance difference image is displayed, subtracting the previous time step’s radiance from that of the current time, using the same technique as with the Tonga volcano data.

In retrospect, this result should not be surprising because the density of the atmosphere at the altitude where the meteor vaporized is insufficient to transmit sound waves, so the answer to the thesis question introduced by the blog’s title is “No.” However, interesting topics that could still be investigated with the existing data could include the following.

Can the spectral signature of the meteor’s chemical constituents be extracted from multiple bands, even when combining the lower spatial resolution bands?
If so, is there adequate data to distinguish a carbonaceous chondrite from a siderite?
Can other bolides be found by comparing JPL’s Fireball date and location data with archival satellite data, and if so, is there a sufficient number and variety that an AI model can be trained to automatically locate unreported meteors in archival or even near-real-time imagery?
What can be learned by studying the spectral features or projected motion of the shadow cast by the vapor trail?

Access New Machine Learning Capabilities in ENVI Student Spotlight Spring 2023