Static Methods On Rugged Topography: Investigation And Application

Numerous static correction methods for reflection seismic exploration have beendeveloped since digital processing method was introduced into seismic data processing,which firstly occurred at1960s. Although in today static correction remains a challengein current seismic data processing, especially in those areas with complex topography.These difficulties are mainly classified as follows：1. It’s hard to keep accurate first arrival picking due to rapidly change in headwave energy, waveform and phase consistence in those complex zones, whilstoccasionally impacted by the presence of strong near surface noises. Additionally,with huge data volume,at present, the efficiency has taken the first place intoconsideration for picking algorithms. Most current auto-pick schemes confrontdifficulty either in precision or efficiency when dealing with these circumstances.2. Refraction static requires a horizontal refraction interface and no lateral velocityvariation, with a surface consistent assumption. These conditions are not satisfied withrugged topography and strong variations in thickness and velocities in low velocitylayers in complex surveys. Reflection hyperbola of seismic data distortions from thesesurveys make inconsistent event to stack, finally influence the imaging quality. Mostexperiential scholars in seismic survey domain believe that refraction methods werepoor in constructing complex near surface modes. However, we will show in ourresearch the ability to build a complex near surface velocity mode by using a refractionmethod.3. Tomography has been used to invert near earth surface velocity model forarbitrary complex underground media and is considered as an idea way to resolve thestatic problem in complex surveys. But drawbacks such as uncertainties in tomography solution and huge amount of computation cost as well as first break picking errorsmake it hard to use, thus impede its industrial application.4. Most current refraction solutions utilize surface consistent hypothesis, which donot always coincide with the real cases. Therefore errors are not avoidable in laterseismic data processing. Physical measurement errors aroused in field geometry alsoinvolved in static estimation. Moreover, more numerical errors are accumulated duringnear surface velocity inversion and static computation when handled by various staticmethods, thus impose bad influence on imaging qualities. So, how to choose properstatic methods to limit static error propagation is an important issue and has beencarefully investigated in our dissertation.As is well known it is very hard to overcome those cumbersome circumstances.As a result we should never trap ourselves into a certain kind of static method. In fact,in the field investigation we need to examine the near surface conditions carefully andcollect as more information as required, and must select a composition of proper staticmethods in later data processing procedures. These works help us to lessen errors tolowest level thus attain a better static solution. In my thesis I probe a series of factorsthat influenced by the complex near-surface static problems includes: static estimatingmethods used to find out static solutions, constrained first wave auto-pick method and acombined energy ratio and normalized auto-correlation first break auto-pick method toimprove picking precision and efficiency, study concerning how to construct complexnear surface models by using refraction waves, and error analyze method to cut downerrors to the lowest level, and a minimum or maximum phase deconvolution operatoradjustment method to promote the static effect.Innovations In my thesis are listed as follows: First, I investigate several staticeffect measurement methods，including static curve display, fitting error display,difference display before and after static application and stacking profile display toanalyze static correction effectiveness, and to help users to determine a better staticmethod; Second, I propose a new first arrival auto-pick scheme based onsurface-consistent constraints, and also a new auto-pick scheme containing acombination of energy ratio of envelopes and normalized auto-correlations. Theconstrained surface-consistent auto-pick method utilize the accurate picking of the seedshot to extrapolate and limit other non-seed shots. However, the new energy ratioauto-pick scheme combined the advantage of the conventional energy ratio auto-pickmethod and the auto-correlation auto-pick method to improve auto-pick precisions. Mytwo methods can tolerate more noisy seismic data, and reduce manual works significantly; Third, I suggest a new approach for constructing complex near-surfacemodels by using refraction method, and also comment some relative static issues. Ianalyze fundamental principles of refraction statics and dispose some problems due toinsufficient acknowledgement of the refraction. I categorize complex topography intofour types: surface undulating, random surface undulating, refractor undulating andrandom refractor undulating, and some theoretical derivation and field data applicationalso illustrated. A further discussion focus on how to satisfy lateral velocity variationswhen build the shallow model. Fourth, I derive travel time tomography statics formulaand wave equation datum correction formula theoretically, and also conduct somenumerical experiments, in order to offer explanation about their outperform thanconventional static correction methods. I also investigate how to eliminate errors invelocity model calculation and static estimations. Through the detailed comparison ofthe rough and precise strip weathering layers, I proved that partial-filling static only issuitable for vertical rays. In case of non-vertical ray occasions WED should be usedinstead; Fifth, I introduce a new kind of static method, named waveform time-shiftcomputation in first arrival wave statics. Here, it is the first to be introduced in currentrefraction statics industry. Waveform time-shift computation involves using a segmentof the first arrival waveform to compute the waveform time, in order to replace directtime algebraic operation, and the result is also a small segment of waveforms; Finally, Ishow a static technique based on the adjustment of minimum and maximum phasecorrection in deconvolution. In my thesis, I estimate weather it was minimum ormaximum phase in current deconvolution operator, and choose the right operator toimplement, and also adjust the phase of the deconvolution result in case of themaximum phase. The phase adjustment improve both the seismic data resolution andthe S/N ratios.