Reply by the author to the discussion by Hualin Zeng and Deshu Xu

I am gratified by the interest of Hualin Zeng and Deshu Xu in my work on preferential continuation. I should mention that the method owes its origin to concepts related to the subject of Wiener regional‐residual separation filtering — ideas and theory that were first introduced to me years ago by Richard O. Hansen.

On: “Preferential continuation for potential‐field anomaly enhancement” (R. Pawlowski, GEOPHYSICS, 60, 390–398).

Pawlowski (1995) presents an excellent method for preferential continuation for potential‐field anomaly enhancement, and it is an appreciated attempt to solve a very common problem in gravity separation. Upward continuation of a gravity anomaly is a very common method for regional‐residual separation in China. One of the main problems in the conventional upward continuation is that it overattenuates regional‐field or useful long‐wavelength information due to deep sources. Sometimes attenuated‐upward continuation of an observed anomaly to a height has to be regarded as the original regional field at the observation plane in order to use gravity inversion to map deep interfaces such as the Moho. Application of the preferential continuation to gravity anomaly in some areas in China has shown very good effectiveness in solving the above problem (Xu and Zeng, 2000).

Preferential continuation for potential‐field anomaly enhancement

A new class of filter transfer function derived from Wiener filter and Green’s equivalent layer principles is presented for upward and downward‐continuation enhancement of potential‐field data. The newly developed transfer function is called the preferential continuation operator. In contrast to the conventional continuation operator, the preferential continuation operator possesses a continuation response that acts preferentially upon a specific band of the observed potential field’s Fourier amplitude spectrum. The transfer function response approaches the response of an all‐pass filter away from this band. This response characteristic is useful for at least two common potential‐field signal enhancement applications. First, it is possible with preferential upward continuation to attenuate shallow‐source, short‐wavelength, potential‐field signals while minimally attenuating deep‐source, long wavelength signals (as often happens after application of conventional upward continuation) Second, it is possible with preferential downward continuation to enhance deep‐source, long wavelength signals without overamplifying shallow‐source, short‐wavelength signals (as often happens after application of conventional downward continuation) Preferential continuation, used qualitatively for anomaly enhancement, ably overcomes these two limitations of conventional continuation enhancement.