Seismic Reflection for Mineral Exploration
June 22, 2020
A Geometrics Geode with ability to record up to 65,536 samples per record allows high resolution data to map stratigraphy and structure to depths up to 2 km using a trailer-mounted 800 lb weight-drop system. A ‘seismic gun’ is sufficient for shallower investigation depths.
Based on the table below, assuming a relatively slow velocity of 4000 m/ns, the straight 2-way travel time to reflect against features ~2 km down (~4km travel) is 1 second. A longer 4 second to 8 second window is sufficient to capture responses due to thicker and slower overburden, for example.
Velocity of Common Rock Types
Rock: Velocity (m/s)
Consolidated Sandstone: 5800
Unconsolidated Sandstone : 4600 – 5200
Shale: 1800 – 4900
Limestone: 5800 – 6400
Dolomite: 6400 – 7300
Anhydrite: 6100 | 20000
Granite: 5800 – 6100
Gabbro: 7200
The results also depend on the geology (e.g., velocity contrasts) and more importantly the overburden which is why it can be difficult to guarantee we can reach the same depths for meaningful data across the survey area. In some areas we will get better/deeper-looking data than others. In some areas we may only get shallow or even no information.
In the following screen shot, an example of the raw data (pre-filtered/processed) showing mostly surface waves and first arrivals with a 10-metre geophone spacing and 24 geophones to half a second. Another proposed spread would have twice the number of receivers and longer depth window (e.g., 4-8 seconds).
For a ‘moving line setup’. The pre-stack filtering steps would include:
- Energy normalization: to prepare for the elimination of surface waves (e.g., remove effects of wavefront divergence and damping).
- Elimination of Surface Waves: These are the high amplitude, lower frequency data that are crucial for MASW work but need to be minimized for reflection work. Includes FK-filter and band pass filtering where necessary.
The following example is what the data should look like after this pre-stack filtering process on the right:
Then the data are ready for ‘semblance analysis’ which will use some refraction interpretation results for better control. This process is interactive. The following example is from shallow data but same process applies to the deeper looking data. Its done for every shot. Then a NMO-stack to bring everything together and finally migration and time-depth conversion to get your typical depth-section. Probably some other filters as required.
Passive surveys, where no active shots are required, can also be considered where many geophones are placed across the property in both dimensions, and micro-tremor data are acquired over a period of time. Colour-contour plan maps for various depth sections are often sufficiently detailed for broad scale exploration.