4 ms·

A lot of data that’s probably already been thoroughly poured over. If you’re searching for logs on a producing well it might be useful. But, I somewhat doubt

A lot of data that’s probably already been thoroughly poured over. If you’re searching for logs on a producing well it might be useful. But, I somewhat doubt the seismic data has much value if they’re giving it away for free.

doodlebugging 8y ago
I have to disagree strongly. The seismic data is probably at least as valuable as the well data. There is a reason that the acquisition contractors and other data owners don't make regular releases of old, legacy data for the public domain. It really has an infinite shelf life due to the processor's ability to periodically employ newer, faster algorithms on newer, faster hardware to produce products that, though they frequently show only marginal differences, are still marketable as new products or upgrades over old datasets. Legacy data, due to the acquisition methods employed decades ago, cannot be replicated today. You will not get a permit to acquire airgun data today using the same energies they routinely used a few decades ago nor will you be able to use a broadband source like dynamite offshore. It was routine back in the 60's and 70's. The bandwidth of data today is different. New source types and improvements to old sources can help but the old data has tremendous value as a calibration. The penetration of energy for imaging the deepest events in the subsurface is so much better in old data due to the low frequency penetration characteristics of old sources (higher energy sources). If you look around for free or publicly available seismic data there really isn't much and ten years ago there was almost none that was easy to find. Industry groups hoping to help newcomers learn by processing raw field data have always been beggars to the data holders. Licensing restrictions follow data everywhere and a lot of it comes with tight constraints on how it can be used and whether it can be published. Most contractors hold tightly to their data because it doesn't matter how old it is, you can always squeeze it through another processing flow and output a brand new, improved product and offer that for sale to your existing and prospective clients. Old surveys get new names, they are merged with new data using match filters and cross-correlations and tied so that it is not possible to tell where the old data coverage ended and the new data began. I started processing almost 30 years ago. Some of the data we processed then was already 20 years old. It served to help a client decide whether a new survey would offer any value to their exploration efforts by contributing a more detailed subsurface image. 2D was a great reconnaissance tool and still is today. By reprocessing some 2D data a client can focus their 3D efforts on proscpects where the potential for success are highest thus cutting their costs and if you think cost cutting isn't a thing in the oil patch, seismic data processing is a loss leader for the big contractors. I did some 4D seismic processing which involves acquisition of new data using the exact same parameters and processing flows as were used in the first survey. New data is then compared to older data so that operators can see the tell-tale changes in their reservoirs which indicate migration of fluids in the subsurface during production or fluid invasion during waterfloods or CO2 floods. Old data never dies nor does it lose its value. Like I mentioned above, new data gets matched to old, old data gets matched to new. Any time a survey is acquired for the first time in an unexplored area, that survey data becomes the ground truth dataset. All future data will be compared to it for quality, bandwitdh, signal to noise ratio, etc. Geophysics, or seismic data processing, really is a "what do you want it to look like" operation. Once you know the acquisition geometry then you can determine everything you need to know to image the subsurface just by smashing it through enough algorithms to filter out all the geology-related attributes like formation velocities, amplitudes - especially anomalous amplitudes, formation thicknesses and their depths below the surface or the seafloor, fluid content, etc. It really is amazing what you can discover without once touching a rock today. It goes so far beyond imaging subsurface structures. I love this field of work. I am pretty happy to see seismic data released for personal use. I will be digging through this to see what I can find.
- convivialdingo 8y ago
  You make some great points. Certainly it’s useful for lots of other cases - I was just considering it from an E&P perspective. They already have numerous surveys and lines throughout those areas. My first job out of college I processed and archived several warehouses of data going back to some of the very first analog signals recorded, TI’s first digital tapes (named GSI at the time) and also digitized paper records from the 1920s. Great first job that exposed me to massive data and algorithms!
  - doodlebugging 8y ago
    Thanks. I would've loved to work with that ancient data. I think one of the biggest challenges as a processor is to understand all the information in the observe reports and how it relates to handling things like field source and receiver geometry, noise issues during acquisition, etc. I imagine that the oldest stuff was treated like a science project with everything notable recorded in detail for future reference. I know that by the 1960's field crews had pretty standard notes they would take. Later in the 70's, as activity picked up before the big bust in 81-82 things became almost sloppy so that it took a lot more effort to figure out what actually happened during acquisition. I worked on field crews back then and one of our observers would fill out his paperwork the night before while he had a few beers and smoked all kinds of things. Then the next day he would just make quick notes if something ended up different. Too bad that when wasn't diligent at modifying his pre-written reports. The prevailing belief on the field crews was that someone would figure it all out in processing later so if they didn't get it right during acquisition we could always fix it. Some of the best projects I worked on as a processor involved unraveling the chains of errors in documentation to improve imaging of old surveys.

doodlebugging 8y ago

I have to disagree strongly. The seismic data is probably at least as valuable as the well data. There is a reason that the acquisition contractors and other data owners don't make regular releases of old, legacy data for the public domain. It really has an infinite shelf life due to the processor's ability to periodically employ newer, faster algorithms on newer, faster hardware to produce products that, though they frequently show only marginal differences, are still marketable as new products or upgrades over old datasets. Legacy data, due to the acquisition methods employed decades ago, cannot be replicated today. You will not get a permit to acquire airgun data today using the same energies they routinely used a few decades ago nor will you be able to use a broadband source like dynamite offshore. It was routine back in the 60's and 70's. The bandwidth of data today is different. New source types and improvements to old sources can help but the old data has tremendous value as a calibration. The penetration of energy for imaging the deepest events in the subsurface is so much better in old data due to the low frequency penetration characteristics of old sources (higher energy sources). If you look around for free or publicly available seismic data there really isn't much and ten years ago there was almost none that was easy to find. Industry groups hoping to help newcomers learn by processing raw field data have always been beggars to the data holders. Licensing restrictions follow data everywhere and a lot of it comes with tight constraints on how it can be used and whether it can be published. Most contractors hold tightly to their data because it doesn't matter how old it is, you can always squeeze it through another processing flow and output a brand new, improved product and offer that for sale to your existing and prospective clients. Old surveys get new names, they are merged with new data using match filters and cross-correlations and tied so that it is not possible to tell where the old data coverage ended and the new data began. I started processing almost 30 years ago. Some of the data we processed then was already 20 years old. It served to help a client decide whether a new survey would offer any value to their exploration efforts by contributing a more detailed subsurface image. 2D was a great reconnaissance tool and still is today. By reprocessing some 2D data a client can focus their 3D efforts on proscpects where the potential for success are highest thus cutting their costs and if you think cost cutting isn't a thing in the oil patch, seismic data processing is a loss leader for the big contractors. I did some 4D seismic processing which involves acquisition of new data using the exact same parameters and processing flows as were used in the first survey. New data is then compared to older data so that operators can see the tell-tale changes in their reservoirs which indicate migration of fluids in the subsurface during production or fluid invasion during waterfloods or CO2 floods. Old data never dies nor does it lose its value. Like I mentioned above, new data gets matched to old, old data gets matched to new. Any time a survey is acquired for the first time in an unexplored area, that survey data becomes the ground truth dataset. All future data will be compared to it for quality, bandwitdh, signal to noise ratio, etc. Geophysics, or seismic data processing, really is a "what do you want it to look like" operation. Once you know the acquisition geometry then you can determine everything you need to know to image the subsurface just by smashing it through enough algorithms to filter out all the geology-related attributes like formation velocities, amplitudes - especially anomalous amplitudes, formation thicknesses and their depths below the surface or the seafloor, fluid content, etc. It really is amazing what you can discover without once touching a rock today. It goes so far beyond imaging subsurface structures. I love this field of work. I am pretty happy to see seismic data released for personal use. I will be digging through this to see what I can find.

convivialdingo 8y ago
You make some great points. Certainly it’s useful for lots of other cases - I was just considering it from an E&P perspective. They already have numerous surveys and lines throughout those areas. My first job out of college I processed and archived several warehouses of data going back to some of the very first analog signals recorded, TI’s first digital tapes (named GSI at the time) and also digitized paper records from the 1920s. Great first job that exposed me to massive data and algorithms!
- doodlebugging 8y ago
  Thanks. I would've loved to work with that ancient data. I think one of the biggest challenges as a processor is to understand all the information in the observe reports and how it relates to handling things like field source and receiver geometry, noise issues during acquisition, etc. I imagine that the oldest stuff was treated like a science project with everything notable recorded in detail for future reference. I know that by the 1960's field crews had pretty standard notes they would take. Later in the 70's, as activity picked up before the big bust in 81-82 things became almost sloppy so that it took a lot more effort to figure out what actually happened during acquisition. I worked on field crews back then and one of our observers would fill out his paperwork the night before while he had a few beers and smoked all kinds of things. Then the next day he would just make quick notes if something ended up different. Too bad that when wasn't diligent at modifying his pre-written reports. The prevailing belief on the field crews was that someone would figure it all out in processing later so if they didn't get it right during acquisition we could always fix it. Some of the best projects I worked on as a processor involved unraveling the chains of errors in documentation to improve imaging of old surveys.