Knowledge of the noise level and the nature of the noise is critical when processing transient electromagnetic (TEM) sounding data. An inadequate noise estimation may result in an erroneous interpretation. We consider the sources of electromagnetic noise and their influence on TEM soundings. The focus is on the characteristics of the vertical component of the noise for an impulse response system which measures log-gated, gated stacked voltages, induced from a decaying secondary field. A fully automatic, hands-off algorithm for calculation of a noise model is developed and replaces the previous manual and subjective weeding out of poor data which was based on the experience of the interpreter. The method enables an assignment of uncertainty to each delay time data giving a more correct and hence more reliable use of the measurements at late times. The algorithm is developed through a comparison of noise characteristics for simulated white, Gaussian noise and EM noise measurements sampled every 2 μsec. It is tested on data obtained with a receiver which implements the log-gating, gate stacking measurement scheme. This automatic noise estimation has been used in the inversion of well over 4000 TEM soundings. Noise weighted data are inverted and the resulting changes in model parameters are discussed for an example with model data and real noise and a field example. The model from the field data is correlated with a drilling.
No takes yet. Share an insight, caveat, or question.
Munkholm et al. (1996) studied this question.