Quantifying Noise in a Land Seismic Image - Stealth Acquisition Experiment

We wish to quantify the relative contributions of the different components of noise in the land seismic image. We acquired a repeat 2D land seismic line. The first pass used a conventional acquisition technique. The second pass was acquired in exactly the same manner, except that once the vibrator pads were down, the vibrators remained stationary with no pad vibration for the duration of the sweep. We call this “Stealth” acquisition; we record no signal, but we record all the noise that we introduce by the process of acquiring the data (except noise introduced while actually putting signal into the ground). A third acquisition recorded an equivalent time of background noise with no source activity. Analysis of the resulting data shows that we generate a considerable amount of noise simply by the process of acquiring the data, some 10–20 dB above the background noise level below 50 Hz, in the raw data. Our processing workflows can deal with this additional noise very effectively and, in the final image, this is reduced to only a few dB above the background level. While this might still seem considerable, it is out-weighed by the noise introduced when the vibrator trucks shake at full force.