longitude and latitude in SEG Y headers?

Oil and natural gas exploration -- geology and geophysics
Post Reply
Quark Scan Inc.
jeepster
Silver Member
Silver Member
Posts: 36
Joined: Fri Apr 01, 2011 8:57 am
Location: Australia

longitude and latitude in SEG Y headers?

Post by jeepster »

I have some vintage data, which only has limited information: shot coordinates in SEG Y header. According to the trace headers, these XY coordinates are tenths of a second of arc. However, dividing them by 36000 does not give longitude and latitude of the survey area. Interpreting as packed decimal does not yield sensible DDMMSS.S either. If one assumes they are northing and easting instead, you get an average shot interval of 3.861 units. The shot interval is 25 metres (82 feet). So the headers seem to be rubbish. A tenth of a second of latitude corresponds to 10.12 feet, so doesn't fit any permutation.
So I am curious if there was some particular way that lonitude and latitude were encoded in SEG Y.
Last edited by jeepster on Sun May 25, 2014 4:07 am, edited 1 time in total.
Quark Scan Inc. - underground utility locating and geophysical survey services in NJ NY and PA
GuyM
VIP Member
VIP Member
Posts: 658
Joined: Sat Mar 24, 2012 11:35 pm

Re: longitude and latitude in SEG Y headers?

Post by GuyM »

jeepster wrote:I have some vintage data, which only has limited information: shot coordinates in SEG Y header. According to the trace headers, these XY coordinates are tenths of a second of arc. However, dividing them by 36000 does not give longitude and latitude of the survey area. Interpreting as packed decimal does not yield sensible DDMMSS.S either. If one assumes they are northing and easting instead, you get an average shot interval of 3.861 units. The shot interval is 25 metres (42 feet). So the headers seem to be rubbish. A tenth of a second of latitude corresponds to 10.12 feet, so doesn't fit any permutation.
So I am curious if there was some particular way that lonitude and latitude were encoded in SEG Y.
First thing might be to look at the co-ordinate scalar (in bytes 71-72); this might have been used with the X,Y headers and could be throwing things off. A scalar of -10 would mean the data in the X,Y headers needs to be multiplied by 0.1, -100 means x 0.01 and so on.

You could also have any wacky combination of numbers ; Decimal degrees, decimal minutes or decimal seconds, or just a packed decimal number with the co-ordinate scalar being used. Of course while the headers are supposed to be IBM 32-bit integers its worth checking to see if they used something else (say 32-bit IEEE floating point - it happens!)

Was never a fan of lat/longs in the headers - X,Y values forced you to figure out the projection system, but with lat-longs people tend to just guess at the geodetic datum which can make life interesting.
jeepster
Silver Member
Silver Member
Posts: 36
Joined: Fri Apr 01, 2011 8:57 am
Location: Australia

Re: longitude and latitude in SEG Y headers?

Post by jeepster »

I checked those things. The 71-72 scalar is -10, so from that I assumed divide by 10, and the header at 89-90 is 2 which means seconds. And they aren't floating point as the exponent bits are zero.

P.S. I edited original post: 42 feet should have been 82 feet.
GuyM
VIP Member
VIP Member
Posts: 658
Joined: Sat Mar 24, 2012 11:35 pm

Re: longitude and latitude in SEG Y headers?

Post by GuyM »

In that case it sounds like someone got it wrong!

A lot depends on how the SEGY was created and how the navigation data was assigned to the trace headers; lat/longs are unlikely to be used directly during the processing (without conversion, that is) and so the mistake might not have been noticed. You can get all sorts of weird results if the data is read from a separate ASCII file and read/interpolated from an incorrect header, or the navigation is read in from an incorrect line.

If the data came from an archive created at the end of the project then there is a chance the data was assigned into the headers and never properly checked; errors tend to get spotted when people use the information to create plots or upload it into workstations, however if you go back far enough or the data is unprocessed (ie raw shots) this might not have happened at the time the archive was created.

I'd be tempted to try and track down the ASCII navigation from another source, and do the merge again.

You can find all sorts of odd things on older data - came across one SEGY dataset that had a weird pattern of missing files, and the range was too large. There were 2 shots missing every ten, and then a whole block of twenty missing every hundred, and two hundred missing every thousand. Took a few seconds to notice that the "missing" file numbers all would have had an "8" or a "9" in them. When the data had been transcribed to SEGY, they hadn't spotted the file numbers were in Octal...
User avatar
geophix
VIP Member
VIP Member
Posts: 941
Joined: Fri Jul 24, 2009 9:26 am
Website: https://quarkscan.com
Linkedin: https://www.linkedin.com/in/groundpenetratingradar/

Re: longitude and latitude in SEG Y headers?

Post by geophix »

Do you have some example numbers for the source coordinates you read from the file? Maybe it's just the scale number is wrong.
aislam3
Member
Member
Posts: 4
Joined: Tue Nov 16, 2021 3:12 pm

Re: longitude and latitude in SEG Y headers?

Post by aislam3 »

Screenshot from 2021-11-16 13-31-00.png
completely lost here. opendtect can't open with correct geometry.
Can anyone please explain ? thanks
GuyM
VIP Member
VIP Member
Posts: 658
Joined: Sat Mar 24, 2012 11:35 pm

Re: longitude and latitude in SEG Y headers?

Post by GuyM »

Where are you expecting to be in the world? Are those actually lat-longs in deg-min-sec or decimal degrees?
If the answer is Yemen (45W, 15N) then you''ll still need to do some detective work..
aislam3
Member
Member
Posts: 4
Joined: Tue Nov 16, 2021 3:12 pm

Re: longitude and latitude in SEG Y headers?

Post by aislam3 »

GuyM wrote: Tue Nov 16, 2021 11:10 pm Where are you expecting to be in the world? Are those actually lat-longs in deg-min-sec or decimal degrees?
If the answer is Yemen (45W, 15N) then you''ll still need to do some detective work..
no this is located in the west coast offshore. I'm stuck. my advisor asking for progress everyday!!! :(
aislam3
Member
Member
Posts: 4
Joined: Tue Nov 16, 2021 3:12 pm

Re: longitude and latitude in SEG Y headers?

Post by aislam3 »

However dividing by 3600 or 3600*1000 gives lat long in degree decimals.
GuyM
VIP Member
VIP Member
Posts: 658
Joined: Sat Mar 24, 2012 11:35 pm

Re: longitude and latitude in SEG Y headers?

Post by GuyM »

That's a start.

You can dump the headers to an ASCII file along with CDP or SHOT, ad then use proj to project them to easting/northing for loading.
Only slight wrinkle is you'll have to guess the Geodetic datum, which can give couple of hundred metres error unless it's in a report somewhere.
WGS84 is pretty common (it's what GPS uses) but it might not be that.

Not sure about OpenDtect loading but you can always update the headers with the easting/northing values.

Just watch out for the COORD_SCALAR and make sure its set correctly.
Post Reply
  • Similar Topics
    Replies
    Views
    Last post