In this paper, we focus on new methods to address the spectral phase and amplitude differences between conventional and broadband streamer surveys in a 4D study. Instead of downgrading the broadband monitor data, we perform “full” deghosting, source-side for the broadband data, and both source- and receiver-side for the legacy data. This broadens the spectra of both vintages to an equal bandwidth and removes the differences due to source and receiver depth variations, to immediately produce very good 4D repeatability indicators. We show the deghosted vintages can be used simultaneously in the SRME modeling step to improve signal to noise and to ameliorate offset sampling issues. We see better SRME results by using this 4D modeling technique which can be very important for 4D’s where multiples are issues.
Recently, the term “broadband” has been assigned generically to a variety of acquisition (and processing) techniques that now routinely obtain better low and high frequency response as well as higher signal-to-noise ratios. A new broadband seismic (Monitor) survey was recently acquired in the Exmouth Basin, offshore Western Australia over a pre-existing, exploratory-style legacy (Base) survey. It is well-known that optimal time-lapse (4D) monitoring is best produced by carefully controlled acquisition repeatability (Lumley, 2011; Johnston, 2013). The new survey has tighter source spacing and streamer separation, which makes it ideal as a base survey for future time-lapse studies, but less so as a monitor survey to the conventionally-acquired base. The effects of the surveys’ source and receiver positional differences can be partially mitigated through a variety of processing techniques, including 4D binning, and/or 4D or 5D interpolation (Trad, 2009). These considerations mostly focus on trying to minimize raypath differences between the base and monitor(s). Although this topic is very critical for 4D imaging, it will not be discussed here. Instead, we will focus on new methods to address the phase and spectral differences from source and receiver depth variations (and their resultant directional ghosts), and how they affect the merging of surveys (and in particular, time-lapse surveys).