Efficient magnetotelluric data acquisition using irregular stations and compressive sensing reconstruction

Abstract

A major challenge in magnetotelluric (MT) surveys is the limited number of stations that can be collected due to the cost of receiver instrumentation and the time required to deploy them in large numbers. Many equal-distance acquisition grid used in MT surveys are either over sampling or faced with severe aliasing. We have applied the compressive sensing theory to MT acquisition and developed an efficient acquisition approach that uses optimized station layout with sparse irregular patterns. This acquisition strategy uses the optimized survey designs to acquire a sparse data set first, and then applies a compressive sensing reconstruction to obtain the equivalent dense data set on a regularly spaced grid. Once reconstructed, the dense data can be used in the traditional manner in the subsequent steps of processing, inversion, and interpretation. Our simulations show that comparable information can be obtained as that from a full regular grid in exploration settings by using as fewer as 25% stations. Furthermore, redeploying the same number of stations in a coarse grid used crustal studies can collect much more information. This new acquisition approach can lead to significant savings in both the acquisition cost and operational time and to maximized information from a given number of stations. The substantial savings in the cost, or the increase in the information content, could make many geoscientific investigations viable that might otherwise be impossible due to limited budget.