Applicability of 2-D Time-Lapse High-Resolution Seismic Reflection Approach to Image Natural Salt-Dissolution and Subsidence in Central Kansas and Improved Post-Processed Vibroseis Data Characteristics

Abstract

The effectiveness of 2-D time-lapse imaging for monitoring natural dissolution of the Hutchinson Salt in eastern Reno County, Kansas was shown to be restricted when comparing high-resolution seismic reflection data acquired in 2008 with data acquired in 2001 across an active subsidence feature. Seismic reflection data were acquired in both 2001 and 2008 across an active sinkhole at the intersection of U.S. Highway 50 and Victory Road in Hutchinson, KS as part of a comprehensive study by the Kansas Department of Transportation to assess subsidence risk and prediction. Data were collected and processed as identical as possible on both surveys. Processed results concluded that acquisition line separation must be within 15% of the Fresnel Zone radius at the target horizon when imaging natural dissolution features. Due to the rate of vertical and horizontal change and the lack of subsurface symmetry common to these features, apparent changes in geology could be related to survey line separation and not true geologic changes. The need for higher fidelity vibroseis data has prompted experimentation on the limiting factors; one of those is the accuracy of the groundforce measurement and associated noise threshold and fine sensitivity of source monitoring sensors and wavelet extraction techniques. Vibroseis correlation and deconvolution methods were compared and used to determine the optimal method for processing high frequency vibroseis data taking advantage of groundforce estimates using improved mass and baseplate measurements. The recorded data were correlated and deconvolved with a groundforce calculated using low noise accelerometers located on the mass and baseplate and is superior to traces produced using standard accelerometers and low A/D digital traces commonly used for calculating groundforces. The use of a calculated groundforce from low noise accelerometers for deconvolution provides a preferred alternative to the common approach of cross-correlating vibroseis traces with a generated synthetic sweep.