A major challenge
State processing massive amounts of seismic data under exploration tax credits
A deluge of seismic data has been hitting the office of Alaska’s Division of Oil and Gas as a consequence of exploration tax credits enacted in 2003 by the Alaska Legislature. The legislation, designed both to encourage new oil and gas exploration and to make seismic data available to companies interested in exploring in the state, has motivated exploration efforts. But, as the number of surveys conducted under the terms of the credits has peaked, and as the volume of data obtained from each survey has escalated, division staff have had to deal with the challenge of keeping ahead of the resulting unanticipated data flood.
“This is a challenge on many different levels,” Mark Wiggin, deputy commissioner of the Alaska Department of Natural Resources, told Petroleum News. “It’s literally a wave front of … ever increasing datasets coming in the door.”
Conducting an appraisalWiggin likened the process of receiving the data from a survey to conducting an appraisal on a house prior to making a house purchase. The division must verify that the data from a survey is complete, that the data can be loaded and that the data are usable before the state can issue a tax credit certificate, he explained. In addition, the Alaska Department of Revenue must go through a parallel massive process, validating all of the receipts for payments associated with the survey, tying the receipts back to the specific survey, to verify the size of the credit that is due.
And, ultimately, under the terms of the tax credit rules, after 10 years the data from the survey needs to be made available to the public in some appropriate format.
The catch in the process is the quantity of data involved.
Raw measurementsThe data consists of the raw measurements collected in the field when a survey was carried out. In a survey, seismic sound sources create sound waves, the subsurface echoes of which are recorded using a series of geophones placed along a line or on a grid pattern. For each sound signal, each geophone will record echoes as a seismic trace, digitized into thousands of individual numbers in a digital recording. Each echo represents a single point in the subsurface halfway between the sound source and the geophone. Seismic processing involves combining and filtering the individual traces to construct images of the subsurface.
But with the raw data consisting of a multiplicity of individual traces, collected using multiple seismic shots conducted at multiple locations, and with surveys increasing in size and complexity over the years, the scale of the data involved has grown exponentially since the seismic tax credit program went into operation.
Between 2004 and 2011 a survey might typically involve less than 5 terabytes of data (5,000 gigabytes). But, starting in 2012, the data volumes skyrocketed, said Paul Decker, the division’s resource evaluation manager. A recent survey involved a staggering 277 terabytes of data, he said.
Delivered on computer drivesThe data are delivered to the division in batches at various times on a series of detachable hard computer drives - division staff have to download the data from each drive into the division’s own computer system. Each drive may contain 6 or 7 terabytes of data. And on a large survey up to 100 drives may be required to hold all of the data. Especially given that computer software has to reconcile the bit count for the downloaded data, to ensure data integrity, it may take two days to process a single disk, Decker explained. And issues can arise with having to deal with data in different variants of seismic data formats.
It doesn’t end there.
Quality controlIn the interests of data security, the division makes a duplicate copy of the data as a backup. Division staff also take the uploaded data and transfer it into the division’s own seismic processing system, converting the data into the processing system’s data format. The examination of the data in this system forms an essential component of the data quality control, enabling division staff to verify data completeness, accuracy and self-consistency, Decker explained. The division could also use its own system to process the data as part of its resource evaluation function.
The processing system enables division staff to determine the exact geographic location from which the seismic reflections in each individual seismic trace originated. This location determination serves several purposes. First of all, it becomes possible to determine what percentage of the data was gathered from within an oil and gas unit, an essential parameter in determining what proportion of a survey is eligible for a tax credit. In addition, when the data become available to the public, it will be possible, on request, to deliver just the data for a specific unit or land area, rather than for a complete survey or for a broader area of territory.
Another critically important reason for linking individual raw seismic traces to geographic locations is a need to exclude from the data available to the public any data gathered from privately owned subsurface land where the land owner does not wish the data to be released, Decker explained.
Data copy for exportHaving quality controlled the data and tied the seismic traces back to locations, division staff then have to make another copy of the raw data, excluding from that data any seismic traces that need to be withheld because of land ownership considerations. This third data copy, stored in a standard format and file structure, then becomes the pool of data eventually available to the public.
The complete process for a single survey takes up to a year to complete, a similar timeframe to that required for the Department of Revenue to audit the survey financial data, Decker said.
Alaska’s Division of Geological and Geophysical Services’ Geologic Materials Center is responsible for providing access to publicly available seismic data, although the mechanism for doing this is a work in progress. The eventual concept is something akin to an online shop, where people would be able to order data based on various selection criteria. However, people would have to go to the GMC with a set of hard drives to actually pick the data up, Decker said. And, given constraints on state funding, DNR is working up a fee structure for obtaining data, with the fees used to cover the cost of the data management, he said.
Processing limitationsThe processing that the division has been conducting requires a huge amount of data storage, and it only has one high-speed computer server for conducting the work. But the division has made good progress in its quality control efforts and is now about three-quarters of the way through processing the tax credit applications expected to be submitted, Decker said. Given that the number of these applications is now expected to drop and given current constraints on state expenditure, it is necessary to be cautious about spending money on further computer system upgrades for what will likely prove a temporary bulge in processing, Decker commented.
But the division is beginning to feel time pressure over completing the seismic data management. It has improved the technology that it is using, has significantly improved the efficiency of the data management process and is looking to rearrange some staff assignments, to make more people available to help with the effort, despite reductions in the state administration’s headcount, Decker said.