For these methods to be successful, NMO correction must first be applied to flatten the events. There are several situations where a sufficiently accurate NMO correction is not available, or where one might like to estimate statics before this NMO correction is applied. For example, in array studies where a 2-D array records signals incident on the array over time from different sources Cole (1988), the possibility of recording events with different dips simultaneously means that no moveout correction will flatten all events. In conventional exploration work, the departure from normal moveout that may require DMO or migration may also mean that NMO correction will not adequately flatten the events.
In both of these cases, an alternative procedure is to estimate statics by maximizing the power of local slant stacks. Rather than attempting to flatten events with a moveout correction, use local slant stacks to estimate the dips present in the data. Then by crosscorrelating each trace with a local stack, estimate the time shifts that improve the alignment of events within the local slant stack. This procedure can be applied iteratively until the stack power is maximized.
Local slant stacks have been used in statics estimation by Kirchheimer 1990. Instead of optimizing stack power by crosscorrelations, he obtains time shifts from the local stacks, and then decomposes these time shifts into surface-consistent shot, receiver, and midpoint components.
In this paper, I describe a local slant stack based method for determining statics by stack power optimization, and illustrate it using 2-D and 3-D synthetics, followed by a real 3-D example, quarry blast data from the SEP passive experiment. As implemented here, the method deals only with receiver statics, as appropriate for the types of data I am dealing with. However, an extension to include shot statics should not be difficult.