EAGE Workshop on Quantifying Uncertainty in Depth Imaging

Depth imaging is almost a mandatory to wholly understand and explore a given field. Absorption effects due to shallow gas at reservoir level are common all around the world. This problem will cause un-optimum image together with the depth uncertainties if not address correctly especially at the targeted reservoir level. Tailored processing flow of Q-tomography and Q-PSDM were used to address these issues, compensating for amplitude, frequency and phase distortion in the data. The velocity plays a significant role, thus application of FWI is mandatory for this kind of environment. Pairing these velocity modelling tools with optimum de-ghosting and de-multiple improves imaging of the structure with a potential of accurate analysis underneath the shallow gas. This consequently reduced the depth uncertainties of the target reservoir.

Acceptance by geophysicists to discuss the reliability of the subsurface and O&G reservoir images they produce in terms of uncertainty is a major breakthrough for the industry Quantifying Uncertainty in Depth imaging is then directly linked to Quantifying Confidence in SubSurface and Reservoir models supporting multi million dollars E&P decisions. Uncertainty on seismic processing and imaging uncertainty cannot be properly quantified by geophysical processes solving inverse problems, but only by kriging algorithms specific to seismic processing, solving reservoir estimation problems. Further and deeper integration of Kriging algorithms in geophysical processing and imaging processes will certainly lead to significant optimization of the parametrization of the geophysical processes and to quantified improvement of E&P project performance

Complex overburdens of low velocity channels can obscure depth imaging due to their heterogeneous nature and lateral velocity variation. Such overburden may cause undesired effects in the form of travel-time pull-down (sag) and possible energy absorption. In depth imaging, the velocity model pertaining to such overburden is of vital importance to bring the objective level(s) to the appropriate image and the depth of reflectors. The aim of this work is to quantify the depth uncertainty using stochastic time to depth conversion method that would span the range of uncertainties of the depth maps underneath such complex overburden which shall enable full range of depth realizations to be generated, gross bulk volume (GBV) ranking that better quantifies the range of uncertainties and subsequently impacts hydrocarbon volumetric estimations.

This paper look at the structural uncertainties in a mature producing field in Malay Basin. In general as more development wells are drilled, the structural uncertainties are becoming less, but it still exist due to - time interpretation, marker and the velocity model. Unlike in newly discovered field - associated structural uncertainties for infill well will cause drilling of expensive sidetrack. Therefore, it is equally important to investigate the role of structure uncertainties and seismic imaging which provide the image - in order to optimize further exploitation activities.

The process of optimal data preconditioning is a critical framework shaping optimum depth imaging with “accurate” locations. This process serves as input to velocity model building and its qualitative evolution for each of the processing stages. Detailed attention was undertaken to suppress unwanted energy and latest “noise-controlled” deghosting method was applied to the data, ensuring recovery of amplitude and frequency for better demultiple and imaging target zones. The total wave field was decomposed into specular reflection and diffraction data allowing their separated imaging and multi-term surface multiple estimation towards optimal velocity model estimation (FWI and tomographical velocity updates) and focused depth migrated images. Latest advanced FWI was utilized to facilitate inversions of weaker energy of lower frequencies to be modelled and updated. The final derived velocity model was cross checked and validated with measured velocities from 15+ wells with velocity mistie uncertainties below 10% due to ambiguities of the interpreted horizons. Refined horizon picking is critical for gamma analyses for depth uncertainties along the horizons. With these views, justification of each step in the process contributing to uncertainty between seismic depth images and actual geological markers from wells and most importantly away from wells is clarified and further quantified.

The paper tries to analyzing the uncertainty of FWI models and the ways to mitigate through integration with tomography and geology

Static geological models are made of horizons and faults which define their geometry. Both faults and horizons (in depth) are uncertain objects, which result from a time-to-depth conversion procedure involving variables such as time maps, velocity maps, markers at wells that in turn might all be affected by a given level of uncertainty. Even faults location may be uncertain and impact Gross Rock Volumes (GRV) calculated in the geological model. This paper details methods which allow simultaneous calculation of depth maps and the associated uncertainty. Focus is put on how to quantify the impact on the global uncertainty on GRV of each Time-Depth conversion input parameter uncertainty. The relative impact of each individual source of uncertainty is calculated on a real case study and the quantitative effect of combining the different sources is estimated and analyzed.

This abstract showcased a data driven approach i.e. Q inverted full waveform inversion that utilized dual azimuth seismic datasets. This is a method to address both the imaging and depth uncertainties on subsurface structure beneath the gas cloud. There are three elements were used for quantifying imaging and depth uncertainties such as comparison of residual move-out map, phase consistency sense check from seismic-to-well tie and total depth uncertainty maps by combining both horizon picking and velocity uncertainties maps based on the depth residual at different well locations.

This presentation shows how ExxonMobil and Schlumberger are leveraging WesternGeco’s Seismic Uncertainty Analysis (SUA) technology to address seismic velocity ambiguity and associated quality & positioning of the seismic image, and provide velocity-based geometric realizations to prospect volumes assessment.