Using low frequency seismic
Feature Articles, May  21  2009 (Digital Energy Journal)

- Analysing low frequency seismic waves, which come naturally from deep in the earth’s surface, is developing as a useful way to find oil and gas. Dale Blue and Audius Meskauskas of Spectraseis explain how it is done

Spectraseis - recording low frequency seismic waves to help detect oil and gas in Voitsdorf, Upper Austria

Low Frequency (LF) passive seismic is a fast moving technology. It draws on an analysis of the passive seismic wavefield of the earth around 1-6Hz to identify new attributes indicating the likely presence or absence of hydrocarbons.

LF passive seismic’s light environmental footprint and its ability to reduce risk and costs have made it increasingly popular.

The resulting information can greatly reduce the risk and improve the success rate of costly oil exploration and drilling activities, and improve decision-making throughout the life cycle of exploration and development.

Spectraseis - doing a survey in Libya

Challenges

LF passive seismic comes with a series of unique challenges. Challenges include the need to take account of naturally occurring and human induced noise, very long trace lengths, and the fast moving nature of the technology.

First, LF passive seismic surveys are very different from conventional seismic surveys.

In a typical LF passive seismic survey, each standalone station consists of a single sensor that records the three components (vertical, north-south and east-west components) of low frequency passive seismic signals for 24 hours or more.

Because the anomalies, which are being detected, are so low in amplitude, it is critical to reduce the amount of noise and this needs to be considered from the survey design and planning stage through processing and analysis.

Due to the sensitivity and low S/N (Signal-to-Noise Ratio) of the measurements, the technician must capture information about the surroundings including the soil conditions, photographs of the nearby area, measurement dates and times, and any other factors that may affect the outcome of the measurements and negate noise.

The local time of day that the measurement is recorded must be considered, as human induced noise can vary throughout the day or even throughout the week. Maintaining all of the metadata and photographs associated with a measurement provides an information management challenge that has to be addressed as efficiently and with as little human intervention as possible.

The second challenge is the nature of the data. Whereas conventional seismic surveys tend to generate large numbers of data traces which are closely spaced together and recorded for a relatively short length of time, LF passive seismic contains very few traces, but these traces are recorded for days routinely.

Traces are analyzed separately and in long, synchronized arrays. Both manual and automatic error detection and corrections are applied in the QC process. And due to the comparatively small number of measurements, every single trace is valuable.

Traces may also have individual bad quality ranges that must be excluded from analysis. The result is that large and complex data structures cannot be efficiently handled directly via the file system and instead need comparatively complex database schema.

Other data must also be collected and considered. Any earthquake signal, for example, can be identified and removed so as not to contaminate the data. It is useful to filter the data based on the earthquake’s magnitude and proximity.

Finally, the fast changing nature of the technologies and the new scientific models we are developing to better understand the geophysical principals involved, require a flexible, modular workflow and the ability to design and constantly update the software tools.

From survey planning to acquisition, processing, analysis and the presentation of the final high quality results, the challenge is how to build the necessary software tools to ensure a seamless workflow.

Data acquisition

In order to negate external noise when gathering data, Spectraseis has developed sensor layout patterns which have been designed using a layout generator that enables a geophysicist to create a survey design that is optimized for the exploration objectives in the area.

This design is then imported into standard geographical information software, so specific receiver locations can be adjusted to avoid natural and man-made obstacles and noise sources.

The final survey design is then loaded into a data acquisition management system that allows the operations manager to schedule and plan acquisition activities. The survey planning information is then loaded onto robust handheld devices, running data acquisition software, which are used by the technicians in the field when placing sensors and downloading recorded measurements.

For soil conditions and other parameters, a touch screen contains lists of available values for the technician to select. After the measurement has been recorded, a wireless connection allows fast download of the information from the acquisition instrument to the handheld device so that all of the measurement data and metadata are contained within the device as a single digital package.

After the field measurements are collected, all the data, including photographs and metadata, are quality controlled in the field and then sent to Spectraseis’ office in Zurich for processing and analysis.

The first phase is data characterization where special software tools are used to characterize the data according to time stability, noise, influences of weather.

The next phase is to remove unwanted noise. Broadband noise spikes tend to occur as a result of human actions, such as traffic or industrial activities.

Spectraseis has designed software which is used to mark unwanted time windows in the data signals and allows the data analyst to view both the raw signal in the time domain as well as in a Spectrogram, where the frequency components of the signal can be viewed in relation to time. Spectraseis’ processing software has also been designed to perform basic filtering and processing of the longer traces mentioned earlier.

Our software also automatically downloads publicly available earthquake data. The earthquake viewer shows a timeline of recorded earthquakes, their relative magnitude and the distance from the center of the selected survey. The data analyst can then filter this viewer to only show earthquakes strong enough or close enough to have an effect on local measurements, thereby determining if a specific noise event may have been caused by an earthquake and filtering out the data in the time window as required.

A data analyst can then use this tool to plot a Syncrogram - multiple spectrograms that represent data recorded during the same time window, and identify specific measurement windows to analyze over a geographic area.

This makes it possible to do further analysis on these synchronous measurements and create 2D attribute profiles or grid maps of LF attributes.

Software development for LF passive seismic also supports research into this new technology. One example is Time Reverse modeling (TRM) which takes synchronous recordings from multiple receivers in a selected area and performs reverse modeling on these signals to determine the source location of the LF anomaly.

The goal of TRM is to effectively create a depth image of the potential hydrocarbon reservoir. TRM has been built in such a way that it is easy for a geophysicist to add new modeling techniques and imaging conditions to quickly see the results.

The results can then be immediately integrated with the forward modeling applications that are tailored for the development of further research into this LF passive seismic technology. Integration with operators’ own seismic and exploration development workflows can also follow.

Biography

Dale Blue is the Software Product Manager at Spectraseis. He has 20 years of experience in E&P Information Solutions, Software Product Management, and Workflow Process Consulting, working in several major E&P services companies. Audrius Meskauskas is a Software Team Leader in Spectraseis. He has over 10 years of experience and list of scientific publications from leading the software development side of research related projects, taking this role in Oxford, Manchester, Ulm and other universities.

Spectraseis

<<BACK