IMPLICATIONS OF SPATIAL-FILTERING ON THE RESOLVING POWER OF MAGNETOTELLURIC DATA

The interpretation of magnetotelluric data, using either inverse metho ds or data-filtering techniques, is underlain by the fundamental quest ion of how can surface measurements be combined to best estimate the u nderground resistivity distribution. Existing methods differ from each other in that they apply different combination rules. Deciding what i s optimal necessarily requires a common ground for comparison. In term s of the simplest case: What is the difference between st standard mag netotelluric measurement and a spatially filtered datum? Can they be c ompared? Our approach to the problem is to represent individual measur ements as weighted averages of the underground resistivity distributio n. Applying this representation to linear combinations of measurements allows a direct comparison between processed and unprocessed data. Th e analysis is further extended to include two-dimensional Backus-Gilbe rt resolving kernels for both types of data. Our results indicate that spatial filtering introduces a focusing effect that greatly improves the resolution at depth.