Extracting full-resolution models from seismic data to minimize systematic errors in inversion: Method and examples

Abstract

Creating an accurate subsurface model is paramount to many geophysical and geological workflows. Examples are background models for seismic inversion, rock property models for reservoir characterization, and geological models of depositional elements for seismic morphological interpretation. The standard workflow for creating subsurface models using seismic data is stratal slicing. The stratal slicing approach, however, may break down in the case of complex stratigraphic or tectonic structuring, such as shelf-to-basin clinoforms, delta lobe switching, deep-water channel-fan complexes, and deformation due to salt tectonics. This paper illustrates how the results obtained with high-resolution inversion and the incorporation of a stratigraphically consistent low-frequency model generated through horizon mapping—called the HorizonCube—improves the quality of the estimation of the subsurface parameters in structural complex settings. Using two data examples with different seismic data and geological settings from the North Sea and offshore Brazil, the paper will demonstrate the increased accuracy of the final inversion result using a data-driven HorizonCube.